值得借鉴的老外打理REVOX A77 开盘机的经验

红灯记 · 发表于 2023-9-21 16:58:54

赶快注册并登录访问我们网站，呈现更多精彩内容！

您需要登录才可以下载或查看，没有帐号？注册加入

x

This page has been in the works for over a year and is not intended to be a service manual.  It's some of my experiences encountered while restoring half a dozen A77s. It is by no means a complete restoration process as there's plenty of that available.  Such an endeavour would be tantamount to re-inventing the wheel.
.  It was written with the intent of providing, hopefully, some insights and useful information on the restoration of the Revox A77 tape recorder.  Very little here pertains to the mechanical aspects of the machine as I've not had any problems in that area, with exception of head replacement and brakes.  Some of these machines have no brakes as they have proven to be more of a pain in the butt than their worth, so a method called 'rockin' the reels' is implemented.  That eliminates the possibility of uneven braking resulting in tape damage, a bigger pain in the butt.  Besides, rockin' the reels is cool.

The detailed description of the test equipment used is to lend some credibility or substance to the restoration process.  Many times I've seen such machines for sale and touted by the seller as having been checked, adjusted, refurbished, aligned, rebuilt, whatever by a qualified technician.  However, there was no mention in any of these rants of the heads.  Some did provide close-up photos of the heads, most of which were on their last legs.  The machines I bought all needed heads which I knew from the start and the price was right, less than $200 including the shipping.  The new head sets cost more than the machine, about $285.  Then there's the MRL alignment/calibration tape although that's a one shot deal.

These "TIPS" are not all my own.  When I acquired my first A77, I had not messed with one before.  A lot of info came from such places as AUDIOKARMA.org and TAPEHEADS.net  There are many folks there who have earned and deserve such titles as Wizards and Gurus.  Some are recorder technicians who have repair shops.
Before starting, you should have a service manual.  They can be found on eBay but make sure you get the seller to show the speed control schematic, diagram 10, as there are two configurations (see photo).  Another source is the Studer archive.

LINK: ftp://ftp.studer.ch/Public/Products/Revox/

These pdf files can be downloaded and printed, no charge and no additional fee if you download two or more.

The two speed control configurations are described here.

#1.077.724/729  This has the IC controller, NE555V (NTE955M) which is just a 555 timer and easily obtained for very little if need be.

#1.077.725  has the cylindrical adjustment coil, T201, (see photo)  T-201 is labeled in purple.

In short, replace ALL electrolytics and you might want to include the tantalums.  Then replace all the Rifa caps.  There's also one on the older speed control board which can be replaced without removing the board, which can be a pain.  The three on the tape drive control board also can be replaced without removing the board which can result in breaking the plastic mounting device.  That plastic has become brittle over the span of some 40 years or more.  Just don't press on the board as it has no rear support.  If replacing the 3 or 4 Rifa caps with the board in tyhe machine, turn the machine upside down and block the board from flexing. Also check R-121, R122, and R-123.  With C-113, C-114 and C-115 removed, they can be checked easily with an ohmmeter. If any of those capacitors short, the resistor will cook.

On very early machines, this board may have a fourth such suppression cap with associated series resistor.  This was seen on a machine I have with a 4 digit serial number.  Replace that capacitor also if present and check the associated series resistor.

Use a variable transformer to slowly apply AC power.  This is good practice after any work has been done to the machine.  Check the 27vdc and 21vdc supplies.  If a scope is available, use it.  For the 27VDC supply, there should be around 140 millivolts of ripple in standby and less than 1V in play.  If the ripple is greater than 1V, there's a good possibility the filter caps are failing or failed or something is drawing a little too much current..  These figures were derived from a working machine with 25.74VDC from the 27 volt supply.  The current drains are 140mV (standby) and 0.5 (play).  It's the one with a SIEMENS C-104.  Another machine exhibited 0.8V ripple.

In the event the power supply needs repair, that board has 4 mounting screws, one of which is next to impossible to remove without first removing the right reel motor and/or the brake solenoid.  If you get lucky, it may be removable with the motor still in the machine but the solenoid will have to be removed.  With the solenoid removed, a good pair of long nosed pliers just may be able to remove that screw.  It seems pointless to say don't put it back.  The power supply board will do just fine with 3 screws, even with the speaker amplifiers attached, if present.

If the power supply has a SIEMENS capacitor (the large aluminum can  C-104) it should be OK.  I can't say this for any other  C-104 but I have yet to find a bad one in any of my 6 machines.  The SIEMENS was seen in the 4 digit serial number machine and that cap passed my tests.  It was charged with 35 to 40VDC and held that charge for a few minutes before dropping by a few volts; it took about half an hour to drop to around 20 volts. (no load)

If either of those two caps do have to be replaced, that can be done in a somewhat unorthodox manner; with the power supply in the machine.  Leave the C-104 installed and move the wires to the replacement.  Fasten the replacement to the old one or any place convenient; just make sure the bare wires on the new cap are insulated.  The other cap, C-101 in the 21VDC supply is easier to replace, unorthodoxly.  Just clip the leads near the cap and solder the new one thereto.  Use alligator clips for a heat sink to avoid possibly melting the solder joint on the board.  As for any other component on that power supply board, only one problem was found in 6 machines and that was D-102.  Obtaining an identical replacement is difficult to impossible unless you find a power supply board on eBay and hope it works.  The other option is an NTE5318.  The device will have to be fitted to the board by criss-crossing the leads as the original device pin-out is ~ + ~ - and the replacement's pin-out is  + ~ ~ - The bridge D-101 can be replaced with just about any circular type device with a 1A to 1.5A rating.  These circular devices have all 4 leads on one end making it easy to fit onto the board.  Radio Shack 276-1152 and NTE 5304 are two examples.

I also learned along the way to replace the tantalum caps as many were found to be open.  My first experience with the tantalums was on the second machine I bought; it wouldn't record on one channel.  Tracing a signal from the input and to the record board revealed an open C501.

Just for schitts 'n giggles, several were removed from the boards and checked not only with a capacitance and ESR meter but also with a DC power supply at the rated voltage of the capacitor and many leaked with the full rated voltage applied.  Meters use a very low voltage to check capacitors but they should be checked with at least the voltage applied in the circuit or better yet, their full rated voltage.  This is especially true for electrolytics used as DC supply filters.  The method used there is to charge the cap to its rated voltage with a DC voltmeter across the capacitor.  Once charged, the DC supply is removed and if the capacitor is good, it will hold its voltage.  If it falls more than a few percent per second, it's probably leaking internally.  A good reference is to do this with a known good capacitor and compare the old one against that.  Despite that large electrolytics have a pressure relief plug, they can spit their innards, so wear goggles and stay a few feet away.  I've never seen one explode even when taken to 30% above their rating but schitt happens.  They do have a surge voltage rating which is usually about 30% above their operating voltage.

So, the tantalums were all replaced from this point on.  I know they can be considerably more expensive than their electrolytic counterparts but having to open up a machine to troubleshoot a problem makes that additional expense worth while.

As for NOS, (New Old Stock), metal parts are ok  but be cautious of pinch rolls.  While it may be new and never used, it's still a good 40 or more years old and rubber ages over time.  The plastic or cellophane™ container/bag isn't going to do a damn bit to stop the rubber from aging and getting hard or gooey although they might work as all my machines have original pinch rolls.  It might be better to send the roller to a place that can replace the rubber and some will even resize the bearing if needed.  Those bronze bushings will wear against the stainless steel shaft on which they rotate and become wobbly, causing tracking errors.  I had an idler from a SONY TH630D re-rubbered and it was wobbly on the shaft.  After several attempts to buff it, it never did work very well and I was limited to how much I could decrease its diameter as it would slip between the two shafts , the drive and the driven.  It cost about $45 (incl. shipping) and when received, I was surprised to find the bearing resized.  It works and also, it's quiet.

To find such places, check out TAPEHEADS.NET and AUDIOKARMA.ORG

The CdS (cadmium sulphide) cell in the end of tape sensor is replaceable with a current new product.  The black plastic housing does come apart (carefully).  See following piccies.  The new  cell has changed manufacturers three times to my knowledge as has Allied's stock number but the cell part number has remained the same.  It an NSL-4522.  The manufacturers' names have changed from Silonex to Advanced Photonix and now Luna Optoelectronics.

The link to the latest item supplied by Allied Electronics is    https://www.alliedelec.com/luna- ... -nsl-4522/70136725/

The photocell is 20MW and the lamp is enough to make it conduct.  The circuit description is in section 5.9.1 of the electrical description of the service manual page 23.  The first ones I bought and of which I still have a few are SILONEX and are identical.

The EOT (end of tape) photocell assembly disassembled. The old and a new (not an NSL-4522) CdS sensor are shown. This was my first CdS sensor and it was a tight squeeze in the assembly but it worked.

These CdS cells won't work with an LED.

When setting the record and playback RF traps, variations of 5:1 between left & right channels can be noticed.  From that which I've seen in Tapheads and AudioKarma, this is acceptable.  Of the 6 machines I have, only one comes close to being equal on both channels.  Just set for minimum.  I've found no one in those forums who had an explanation or fix for this.  I've replaced C517 & C808 and checked L--501 & L-801 to no avail.  I don't know if the adjusting cores could be at fault but I've even inserted steel screws into the coils with no luck.  One machine works well with one core removed.  Go figure.

A worthwhile note.  Using an ordinary multimeter may not work if its low range on AC volts isn't sensitive enough.  These voltages are less than 350 millivolts and the minimum can be as low as 20 millivolts.  While no DMM ordinary will give accurate results at 120khz, actual voltages are not required, only a minimum.  So, if your meter will read as low as 20 millivolts AC, you should be ok there.

With a DMM, give the meter time to stabilize after turning the core.  A scope works best as the change is instantaneous.

There are some notes on the heads and bias traps at the bottom of this page in the section on The Heads



Miscellaneous Specifications

Tape speed on all these machines is stable within +/- 3hz at 3khz from an MRL (Magnetic Resource Laboratory) calibration tape.  Specification is +/- 0.2% which, at 3khz, is +/- 6hz.  This applies to 3.75, 7.5 and 15 ips
It is therefore safe to assume that WOW & FLUTTER are very close to spec., +/- 2.4hz and +/- 3hz at 7.5 & 3.75 ips, resp. and TAPE SLIP which spec says is less than 6hz @ 3khz (0.2%)

I have no specifications for the 15 ips machines.

HARMONIC DISTORTION at peak level (+3VU) and 0 VU, resp. is as follows

7.5 ips  2% and 0.6%, resp.       and at       3.75 ips  3% and 1%, resp.

Only the 0 VU measurement was made and all were well below the 0.6 % and just under the1% after subtracting the generator's THD. Distortion at  peak level wasn't considered due to my recording peak levels never exceeding 0 VU, typically < -3VU.

Frequency Response Specifications
7.5 ips 3.75 ips
30hz - 20khz +2 / -3 dB 30hz - 16khz +2 / -3 dB
50hz - 15khz +/- 1.5 dB 50hz - 10khz +/- 1.5 dB
Actual frequency responses are discussed with the appropriate response curves shown later.

Test Equipment

All this stuff may be old but they are all in specification

DANA Model 8100 frequency counter
The digits shown here are directly from a genny and are to show the nixie display.  It can sample at 0.001 sec to 10 sec.  It is set here at 0.1 sec.

It will display 8 digits in 1 sec and 10 sec sampling rates.

The last digit appears blurred due to its changing 10 times a second; the genny and the counter haven't warmed up enough.

TEKTRONIX 465 squiggle toob
(That's what Paul Klipsch called them)

This was bought new in 1984 and still meets spec.

I guess the name says it all.

The three meters on top are Fluke 8050A models dating to the early 70's.  Purchased on eBay, they all worked  but needed zebra strips to get all the segments of the displays working.  All showed no signs of parts having been replaced on the board.
The center unit in the lower compartment is a Tektronix AA501 distortion analyser with a built in AC voltmeter.  This unit was also purchased in 1984 with the scope as was the FG-502 function generator (left)  On the right is a DM-501 purchased years later. It was repaired (flaky switches and old caps) in March, 2014 by Norway Labs at a cost of $347 including shipping both ways.  It's certified to meet original specs which are:

0.05% on DCV; 0.2% DCA; 0.6% ACV (40hz - 10khz) and 1% between 10hz - 40hz and 10khz - 20khz);  0.6% on ACA, 20hz - 10khz).  Ohmmeter is 0.15%.

The DM-501 is used to check the accuracies of the other meters.  The AA-501 specs at 2% (20hz - 20khz) and still meets that.  The Flukes, when new had the best accuracies of 0.03% DCV and 1% ACV (20hz-20khz)  While they don't meet that now, they do meet the specs of the DM-501.  As I mentioned earlier, the name says it all.  Those guys are about 45 years old.

HEATHKIT MODEL IO-1275
Linear/Logarithmic Sweep Generator

This thing has been in my possession since I bought it way back in the late seventies and it still works. This unit, although not good enough to measure THD of an amplifier, has an advantage over the FG-502 in that its output is variable from 1mV to 7V (rms) into >50W, whereas the FG-502's range is 0.354v to 3.54v. This is useful when doing the record head azimuth adjustment which requires 3mV applied to the input. There are several Heathkit pieces here and all working but have been replaced by more versatile equipment.

The computer on the left has the LMS board which fits into an ISA slot, no longer seen on motherboards.  The OS is WIN ME.  It also contains a plethora of BASIC programs I've written many years ago to design crossovers (passive filters), active filters, calculate the manually measured Thiele-Small speaker parameters to name a few.  There is a working spare neatly tucked away.
The one on the right is a WIN XP machine that I used to use on the net.  It was taken from my stash for the sole purpose of running the PICO scope (seen on top).  It's a USB powered scope about the size of a deck of cards.  The scope function is seldom used unless I need to save and print a screen, which is far better than trying to take a piccy of the 465 scope.  It also has digital storage capability great for catching pulses but the main reason I bought it is because  it  has  a  spectrum analyser.    The cost.  $400.  Incredible device.

The little fella in the center is a Brother HL-2040 laser printer used only for text files.  Anything requiring color is saved to a thumb drive and printed on a Canon color ink-jet if required.  Most of the color files are used in web pages.

21878

The number 21878 is the serial number of a half track high speed A77.  It was the first such purchase and restoration so the learning curve was lengthy.  The first (top) chart is an early one run for setting the high frequency bias and equalization adjustments.  This was done with the LMS1.  By feeding the LMS sweep signal into the A77 and monitoring and feeding its output back into the LMS, the following chart is plotted.  These are actual responses of the recorder as seen from the playback head.  Any adjustments made on the record amplifier are simultaneously (actually about a second later) seen as the data is plotted.  The trace stops at 100hz as these adjustments have no effect below 1khz; I just have a thing for decades, hence 100hz.  This chart was run in 2011.  While virtually flat at 7.5 ips from 100hz to 15khz, (blk and red) it does drop by 5 dB at 20khz.  This is still within the +2 to -3dB spec (5dB) but had the sweep gone to 30hz, that 5dB would be more like 8dB to 10dB.
The rise in level with frequency was the best I could get at the time as the original pots were flaky and the 15 ips curves were the most stable.  Just about any other setting on the pots, while somewhat flatter wouldn't hold their position if the record boards were tapped on their ends with a pencil or if the machine was lifted off the bench and placed down.  Very sensitive to vibration.  Yes, the pots were cleaned with alcohol but the wipers couldn't be re-tensioned without risk of breaking.  No disrespect to Revox intended; these pots were not intended to last for 50 years.  With the new pots, the record boards were actually dropped 3 feet to the floor with no change in settings.

This HF rise was not considered a problem as my hearing stops at 13khz. At this point I stopped as it was good enough for the time being.  Another A77 had arrived and I was anxious to get into that one and hopefully learn more about that which makes these things tick.

Four years later, something went awry with the speed control and the machine was placed on the bench again. See next curves

I suppose the trim pots on the record and oscillator boards should all be replaced as they are very touchy, hence the slight rise around 15khz in 15ips mode.  If head azimuth were off, I would expect the HF response to drop.  On the other hand, it is possible the azimuth is slightly off and the rise is due to over compensation and over bias.  But, they sound damn good anyway and without better pots, there's not much point in trying again.  The drop after 15khz at 7.5 ips does seem to indicate a slight azimuth error.

21878

These curves (below) were run in 2015 after the speed fix, which, if memory serves me well, was a failed C213, which is one of those infamous RIFA caps, the problems of which I wasn't aware at the time.  Since the speed was in spec then, well enough was left alone.  A lesson well learned - replace ALL the Rifa caps as well as the FRACO electrolytics.  I also learned along the way to replace the tantalum caps also as many were found to be open.  My first experience with the tantalums was on the second machine I bought; it wouldn't record on one channel.  Tracing a signal from the input and to the record board revealed an open C501.
Just for schitts 'n giggles, several were removed from the boards and checked not only with a capacitance and ESR meter but also with a DC power supply at the rated voltage of the capacitor and many leaked with the full rated voltage applied.  Meters use a very low voltage to check capacitors but they should be checked with at least the voltage applied in the circuit or better yet, their full rated voltage.  This is especially true for electrolytics used as DC supply filters.  The method used there is to charge the cap to its rated voltage with a DC voltmeter across the capacitor.  Once charged, the DC supply is removed and if the capacitor is good, it will hold its voltage.  If it falls more than a few percent per second, it's probably leaking internally.  A good reference is to do this with a known good capacitor and compare the old one against that.  Despite that large electrolytics have a pressure relief plug, they can spit their innards, so wear goggles and stay a few feet away.  I've never seen one explode even when taken to 30% above their rating but schitt happens.  They do have a surge voltage rating which is usually about 30% above their operating voltage.

So, the tantalums were all replaced from this point on.  I know they can be considerably more expensive than their electrolytic counterparts but having to open up a machine to troubleshoot a problem makes that additional expense worth while.

Regarding the curves below.  After the speed problem was resolved, a few minor checks were performed, one including checking the bias and equalization settings.  The original pots were flakier than in 2011, so they were replaced.  Original physical size pots weren't available so smaller sealed pots of higher quality were used.  These had to have their terminals lengthened by 1/4" to fit the board.  AWG20 wire was used to provide stability as thinner wire tended to let the pot move inward when applying slight pressure from the adjusting screwdriver.  These pots were very smooth when checked with an analogue ohmmeter over the complete arc of rotation.  The curves below are the result.

While the curves rise with frequency, keep in mind that the scale is 1dB/div.  Referencing from 1khz, the response is +/- 1.5 dB from 30hz to 20khz.  The overall deviation is within 3dB, which is 2dB less than spec for 7.5 ips machine.  The result in this graph might be expected of a 15 ips machine.

56551

A regular machine, 3.75 and 7.5 ips The red & black traces are not good at all. Q502 on the left channel was replaced and seems to have eliminated all the squiggles seen in the black trace. The green one is that with the replaced Q502. The capacitors, C501 & C502 may account for the 4dB rise in output as no settings were altered. C501and C502 are at the input to the record amplifier. I found one C501 in my used parts box, and checked it with the ESR meter and it showed 7.6 W. A new one is usually less than 2 W, typically around 1 W. I've kept most of the parts from these rebuilds knowing I would eventually get an ESR meter and learn a little more about how component variations affect the circuit. The C502 caps showed no appreciable difference between the old and new. The roll-off at the low end is normal but the 2dB difference between the channels isn't. This is not considered a serious problem because in the initial setup, section 6.3-52, it states to set the balance within 3dB between each channel. I usually set them as close to equal as possible, like 0.2dB but with moving the machine around during repair, this could easily drift. This 0.2 dB is a difference of 5 millivolts. Consider the signal from a pre-amp's tape output which can be around 0.5V, that's 40dB above 0.2V
The high frequency roll-off at 15khz is another story. However, I left it as my hearing doesn't go past 13khz and also the nature of the music I play on these machines. Vintage jazz from the roaring twenties and the dirty thirties, mid 50's to mid 60's rock n roll, doo-wop and rockabilly. That stuff recorded back then is usually deficient in the high end, especially if recording from old vinyl. Also, I record on vintage tapes, if they're not damaged, and those old magnetic media need all the bias they can get. They're probably worn out also from playing on recorders with pressure pads to hold the tape to the head face. Those old tapes can help give a juke box sound, along with certain vintage speakers.

It's all about nostalgia.

87349

Another half track high speed.  This fella needs a little more work, probably on the record boards.  The two pairs of traces were to find out how much the high frequency response may be affected by recording at a higher level.  I remember hearing it said that for best HF response, the recording level should be kept between -10VU and 7VU.  This, of course, demands a low noise tape or Dolby.
Both red and pink traces are right channel; black and blue are left channel.  The differences between left to right is again within the specified 3dB as per section 6.3-52 of the service manual, in this case, 1dB or less.  The level increase was done using the genny output as that would maintain the same relationship between left and right since the genny output is split between both channels with a Y connector..  As can be seen, the difference increased by about a half a dB in the blue and pink curves.  The overall shape of the curves at both levels does seem to be very similar, even above 15khz.

It should be noted though, that these two pairs were recorded at -7VU and -4VU.  It also should be noted also that there's a typo in the MAP section at the bottom of the graph regarding curve 19 which states the genny output was 105mV.  That only 0.424dB above 100mV.  The correct figure is as stated in curve 20, being 150mV which is 3.522dB.

157850

This is a quarter track, 3.75 & 7.5 ips machine.
These curves are a comparison among various tapes.  Note the top four curves; purple(40), black(16), turquoise(41) and black again(47), from top to bottom. All within 3dB, 2dB better than the 5dB spec. With Quantegy 632 (16 upper black) and RMG LPR35 (40 purple).

All curves are run with the same input level and the same bias and EQ settings; nothing but the tape was changed and all traces are run in the middle of the tape on 7" reels.

In the MAP section at the bottom of the graph, ref curve 16: the JFSAG means Just For Schitts and Giggles.  This was because I wanted all curves run at the same time.  There was an earlier curve of unknown date.  All are run on the left channel.

The 7"R and 7"T mean 7" Reel set at 7" Tension.

The lightest green(42) is  Scotch 207 that I bought around 2012 on eBay.  It was an unopened carton of 10 reels.  Not bad stuff.  A minor bias and/or EQ tweek could give a flat response.

The slightly darker green(43 bottom) is very old and highly used Scotch 150.  The dark green(44 middle) is also Scotch 150 but with IEC equalization.  Quite a difference, almost +4dB @ 10khz.

These three curves were run to see if there was any difference on frequency response with different tension settings, which loads the reel motors from 42 vac for 7" reels to 55 vac for 10" reels.
The black curve (16) is the same as that in the previous set of graphs; 7" reel and 7" tension.  The green curve (12) is the same reel and tape but with the tension set for 10" reels.  Note the difference above about 13khz, the irregular trace and the level drop below 2500hz.  I pondered the reason for an unusual 42V instead of 40 but I suppose Revox engineers played around with these voltage levels in search of the best level for optimum tape tension.  It probably doesn't matter much on the right reel as it just takes up the tape.  A higher voltage would be required due to higher mass and decreasing mechanical advantage as the right reel diameter increases.  The capstan pinch roll, when engaged will block the tension on the tape from the heads but the tension on the left reel WILL affect the intimate tape to head contact.

The more expensive later microprocessor controlled machines of most, if not all, manufacturers used variable sensors on the left and right tape tension arms and dual capstans which could maintain a very uniform tension regardless of the diameter of the tape on the  reels.  They also maintained tension in FF and REW as well as function as an end of tape switch and a real time counter.

The blue curve is #16 again but with IEC equalization.  This comes in handy when recording from worn vinyl.  Brings out such things as cymbals and brushes  and brass winds (sax) from barely audible to nice, real nice.

If you, who are reading this, haven't heard rock n roll and rockabilly from the late fifties, you really missed something.  If I may suggest, get yer butts on You Toob cuz you'll never hear the likes of that again.  What an era.

These three sets of curves were run 2 days after the set above. Probably a last check prior to replacing the cabinet.
All are run with Quantegy 632.

Full Track

This is a half track 3.75 and 7.5 ips machine, one of my last eBay acquisitions bought essentially as a hangar queen.  However, since I didn't need any parts after a few years, it was put onto the bench.  Prior to that, it was used as a modified fast forward machine to transfer tape from 12 inch pancakes (5000 feet) to 10 inch reels.  A supply reel table was borrowed from an AKAI GX-4000 and mounted on a heavy board to the left of the A77..  Somewhat of a Rube Goldberg contraption, but it worked.

Verifying the signal being present at the outputs of the input board, it doesn't get to pin 13 of the record board on one channel.  It's possibly a faulty contact in one of the pre-record switches.  Further trouble-shooting at this time was pointless as the intent was to install full track heads on the working channel,.  A full track machine tantalized my curiosity for decades.  It was placed on my bucket list.

The full track heads were obtained from eBay and are genuine Revox heads; they cost me a little over $300 with the shipping.  While it has been suggested I use this machine to make azimuth calibration tapes, that would be based on a lot of assumptions, most of which would be fraught with errors.  It's being done for nostalgic reasons.  Nothing else.

The procedure given by Studer is quite involved.  It suggests removal of the cross-talk circuit since it is no longer needed.  Also, while only one record amplifier is required, both playback amplifiers are retained.  The reason is that one VU meter monitors the incoming record signal while the other monitors the playback signal.  This is standard on later 3 head machines when switched from SOURCE to MONITOR/TAPE but the A77 did not have active VU meters in playback unless it was DOLBY equipped..

If I get this machine working, all trimpots may be be replaced by 10 turn potentiometers and these mounted in a shielded chassis and connected to their respective boards with shielded cables.  The end result will resemble an old AMPEX from the fifties.  I also have 5 inch 200 mA DC meters already modified to work as the original meters. R353 and R354 had to be changed to 11kW. If you're wondering what purpose this will serve, well, let me say this about that It'll look kool.

Some people put fancy paint jobs and fancy wheels on cars which do nothing for the car's performance.

It just looks kool.

I've wondered how the first azimuth alignment tape was made.  I recall reading a paper (wish I could find it) that stated that a signal was recorded to the tape with the gap in the record head as visually perpendicular to the tape as possible.  The tape was then viewed under a powerful microscope and the alignment of the magnetic particles could be seen.

Checking/adjusting head height.  Some prefer a transparent tape.
The red shaft is that of a plastic screwdriver to keep open the playback head shield.

Just a close-up of the full track record (left) and playback (right) heads.
The purple cable tie did a better job that the screwdriver (above) It has less tendency to dislocate itself due to lighter mass. Besides, it's purple.

Erase head (left) and record head (right)

The full track, fully operational but not quite fully assembled. This is due to my getting around to extending the power switch shaft. (next) The switch knob was missing and I ain't paying some greedy individual $19 to $25 for a plastic knob, not including the inflated shipping cost they want. Heck, that knob can be mailed for less than a dollar
Soooo, the next best thing was to modify the switch shaft.

This machine, along with the two AKAI's in the rack and another not seen in this photo are all wired into the home made patch bay seen on the right side of the rack, just above the right reel.

I made it in lieu of using one of two I have so it would match the solid oak rack. It allows 6 recorders to be connected to the TAPE 2 inputs/outputs on the pre-amp. It also allows tape to tape transfer from any combination of recorders to any any combination. Those other numerous features weren't considered at the time and will never be used.

The 3 full track heads. Note that there are only two wires on each instead of 3, 4 and 4. ]

The machine was a low speed unit,
3.75 ips and 7.5 ips half track.

The left channel has too many issues, which is why I bought it. Two previous machines I bought for this conversion fell prey to my just having to mess with the inferior channel and ended up repairing them. They eventually had new heads installed. But, this machine came quite inexpensively, about $110 and without a cabinet. The poor thing was in deplorable condition. My only wish is that it was a high speed, 7.5 ips and 15 ips. But, at least I finally got to use those new full track heads which cost me around $300 for the set. They were obtained separately over a period of a year or so.

POWER SWITCH KNOB SHAFT MODIFICATION

The power switch shaft on this machine has been extended. A steel spacer of length 3/4 inch and `1/4 inch inside diameter was drilled with 4 holes which were then threaded.  Two holes are spaced about 3/8 inch and diametrically opposite two other holes.  Two holes diametrically opposite can be drilled at once.  Depending on the drill size used, they can be threade3d with 6-32 or 4-40 set screws of length 3/16ths inch, the larger size being preferred.
Two screws hold this spacer to the original shaft, screws against the flats.  A steel or aluminum shaft of diameter 1/4 inch and about 3/4 inch long is inserted into the open end and locked in with the other two set screws. A regular knob with a 1/4 inch hole can now be installed.

The hole in the escutcheon plate may have to be enlarged  to about 1/2 inch to clear the spacer.  The knob will cover that.

The spacer, available at Ace Hdwe. and others.
The wood dowel makes it easy to drill. With the spacer in a small groove cut into a piece of wood and the ends of the dowel fastened down in a suitable manner, say with a pair of screws with wide heads or a washer on each end, your hands are kept away from the work. It can be drilled through with a hand drill or better yet, a drill press.

Then the 4 holes are threaded.

Below, the frequency response of the full track on Quantegy 632 tape.  The blue is 3.75 ips, red is 7.5 ips. A few components on the record and playback boards had to be changed and some removed.  That info can be found on the Studer archive site.
ftp://ftp.studer.ch/Public/Produ ... _A77_Mono_Diagr.pdf

The values of some of those resistors that are changed can be slightly changed from that given to adjust the response curve.  That's how these curves were obtained to be flat within 2dB, a lot better than specification, +2 to -3 @ 7.5 ips, 30 to 20k  This machine, at 7.5 ips is flat within 1.5dB from 20 to 20k and at 3.75 ips is flat within 1.0dB from 30 to 16k.

So, how much better will this one sound over one just within spec?  Probably not detectable.  The main and only reason for doing it was to find out if I could.

5308  Little Jewel

A Real Oldie  SN: G005308
I found this jewel on eBay on May 9, 2011.  It came from Alberta, Canada.  When I first saw it, the auction was almost over and no bidders.  The starting bid was $100 so I tossed in a bid of $120, just to get instant notification of any higher bids.  I was prepared to go higher as I was looking for a recorder rather than a deck.  No bids were placed so I got it for $100 plus $60 for shipping.  It was rated high by the seller but I took that with the proverbial grain of salt. I was very much surprised upon receipt of the machine.

First, it works, even the brakes.  All lamps work but only one of the VU meters.  Turned out that C507 was shorted; both were replaced on each board. The cabinet is pristine and when I removed the cabinet, no dirt or dead bugs inside; it looked new.  The heads show only a shiny area where contacted by tape, no actual wear.

The seller was contacted shortly after it arrived and I was told that his dad bought it in the 60's and hardly used it. It was stored most of its life in a closet in the house; his dad was a traveling salesman.  It never saw a repair shop so all internal adjustments are original, hence none made by me.  After telling him a little about my "thing" for the A77, he said he was happy knowing it found a good home.

It lay around my collection for years and I played it a few times, showing it to a few friends.  It was finally placed on the bench for investigation due to its age and originality.  By then, I was very well aware of those Rifa caps and wasn't about to push my luck any further than I already had.

Tests were performed but again, no adjustments made as everything was in order with exception of the speed which was a little slow.  So, I had to remove the power supply as trying to get at those two adjustments is more difficult that removing the power supply.  Besides, there are old electrolytics that HAD to be checked.  Q209 is the old TIP-27 in a TO-220 case.  Despite that the speed was easily adjusted correct, C213, a Rifa and C212 were replaced, the latter checked as questionable and due to the difficulty of access, it was replaced.

In the end, a few tantalums and the speaker amplifiers'  electrolytics were replaced which reduced ripple and hum and ALL Rifa caps replaced.  This machine has 4 such caps on the tape drive control board; later models have only 3.  After restoring 5 units, I was familiar with which tantalums to look at.

The heads were in good alignment and not touched as the red seal glue (DFW - Don't F+++ With) was still intact.

Three views of the power supply with the speed control board attached. All pertinent connections to these boards are connected with extender cables I made, which can be seen easily in the first two piccies at bottom center. These extender cables were made with the same colour wire as the originals to avoid confusion. Where a tracer wire was needed, a white wire was coloured with felt tip markers.
Where these extender wires are connected to those coming from the machine, the connections were covered with shrink tubing using a size tubing that would hold its position without applying heat.

These extender wires and shrink tubings are kept in a plastic bag for future use.

Probably needless to say but I'll say it anyway, stability of the assembly is verified and all unnecessary paraphernalia is removed from the area.
A note on removing the power supply.

There's a very thin black wire connected to the low voltage power supply near the lower left mounting stud to a lone foil on the board. I've seen some of these wires broken and it took a while to figure out where that thin wire went as it broke on the foil.

Inspection of another machine solved that.

A piccy of that wire's location is next.

The Thin Black Wire - All A77's
Sometimes broken unknowingly during a previous disassembly.  This wire comes from the power transformer and probably connects the transformer's laminations (core) to the machine chassis.  It's shown left of  section D of Diag, 2 just under the two secondary windings that go to the power amplifiers' (optional) dc supplies.

It's shown as a 'ground' but is not an earth ground, since this machine has a two prong AC plug.  Neither is it a common line for DC, although the manual does refer DC common as ground and the chassis will give accurate DC readings when such are measured to the chassis.

The low voltage power supplies show their DC common lines as floating.

Chances are the machine will function without that wire connected but if a mild hum is noticed, check that wire's connection.

5308  Little Jewel

These next three charts were obtained about 4 years after purchase.  No adjustments were made to bias and EQ.  Assuming all settings to be as done at the factory, the characteristic rise in the high frequencies may be somewhat normal and/or due to aging components as all are well over 40 years old. The bias and EQ settings are all original and probably factory.

The top chart shows the response using Quantegy(Ampex) 632, new tape bought on eBay in 12 inch pancakes and in an unopened carton.  These response curves exceed the original; Revox specs.  I'd love to know what made these differences so they could be duplicated, maybe.

The second chart shows the response curves with different tapes, as indicated.  All are run on the left channel and at 7.5 ips.  Note the red curve, RMG-LPR35 with it's incredibly flat response.  The 4dB gain is due to the tape being "high output".

The third chart shows response curves using some very old tapes.  The TDK GX35 which I bought in '84 is close to identical to the older 632.  The Scotch 111 and 141 performed better than expected, especially considering their age and condition.  The Scotch 190 and Ampex 611(new in boxes) failed drastically but in all fairness, I don't doubt their performance could be improved with bias & EQ changes.  The Ampex 611, of which I have several NIB reels doesn't do well on any machine biased for Quantegy 632.  There is one machine biased for very old tapes, the Ampex 611 is one.  Some other very old and NIB tapes do perform well on that machine.

It should be noted that once the machine is biased for a tape, that tape will perform well on other machines in playback.  The bias and EQ settings are on the record boards and affect the recording signal more.  However, since the tape is monitored through the reproduce amplifiers, there respective response characteristics will have an effect on the bias and EQ settings be4ing made although I would think them to be small.  If one takes the time to perform the Frequency Characteristic, Playback  Section 6.3.4 in the setup procedure.  A DMM isn't going to work as it's too frequency sensitive, especially above 5khz and that's with a damn good DMM.  A DMM or bench meter capable of frequency insensitivity  beyond 20khz will cost several hundred to over a thousand dollars.

45532

This is the machine mentioned earlier that was to be converted to a full track, however, Murphy changed that.  Since it already has new heads, I ran another frequency response and as expected, these curves were much the same as those derived several months ago in February, 2017.  A little more tweeking was done but to no avail.  Then it came to mind that I bought new alignment tapes from MRL and the reason was even more profound.  It was discovered that my TEAK test tape which I bought in 1985 was defective.  Soooo, I checked the heads' alignment with the MRL tapes and sure enough, they were off.  So, I realigned them and went through the whole setup procedure, compensating for the +3dB difference between the 185nW TEAC tape and the 250nW MRL tape. Actually, it's +2.62dB.  The result was astounding.  The machine now meets better than spec by +1dB, for all that's worth.
Not liking to replace the heads with full tracks due to the new performance of the machine, I bought another A77 on the bay.  It's a half track, 3.75/7.5 ips unit.
this is the old frequency responses for 7.5 and 3.75 ips.  This was the reason to convert to full track.  These curves were based on the erroneous assumption that the heads' alignments were good.

After realigning the heads with a new calibration tape, this is the result. With respect to the level at 1khz, the response is better than spec. by 1dB. Although that difference isn't audible, it's still better than spec.
SPEC @ 7.5 ips = +2 / -3 dB 30hz - 20khz (actual -2 at 30hz, +2 at 12khz and -2 at 20khz, or +/-2dB)

and +/- 1.5 dB 50hz - 15khz (actual = -0.5 at 50hz, +1.2 at 12khz and +1 at 20khz, within +/-1.5dB)

I will admit that the above sounds like nit-picking and that the curve doesn't look that flat but the point is, it's in spec. Keep in mind these machines are from the sixties and technology then wasn't what it is today, especially not for a home audio recorder.

First, there is NO IEC equalization for 3.75 ips. The speed selector switch, S6 when in 7.5 ips mode, bypasses resistor R-811. With NAB selected by S2, R-810 is also bypassed. In IEC and 7.5 ips, R-810 is inserted back into the circuit. When in 3.75 ips mode, S6 bypasses R-810 rendering the position of S2 meaningless. R-811 is now in the circuit for low speed EQ.
See Diag. 8 and section 5.6 on pages 20 and 21 for more on the feedback EQ circuit.

The high frequency response at 16khz is below spec at -6dB with respect to 1khz. One or more of the components in the feedback circuit could be out of range. The EQ trimpots, P501 and P502 on the record boards are at minimum and maximum rotation, respectively. Due to the possibility of damaging the foils on the playback boards, attempts at removing components for checking was deemed unworthy. If the response could not be taken into or even close to spec, further diagnoses would have been performed. Also, the lower speed is seldom used for recording or playback and there are 5 other machines here that perform better at the low speed.

Another option is the acquisition of another pair of record and playback boards to rebuild.

Comparison if IEC and NAB responses, 7.5 ips. The pink and black pair are those of the IEC equalization.

The Heads

And now for something completely different
This shows the difference between a quarter track, left and a half track, right.  The colored markings are in proportion to 1/4 and 1/h the tape width, that being 1/4 inch.

The numbers on the left refer to the tracks on the tape, track 1 being on top. The dark bands do the recording or playback..

S1T1L means side 1, track 1, left.  When the tape is flipped to side 2, it becomes side 2, track 4, left.  When  flipped, track 4 and track 2  swap positions with 1 and 3, respectively.

It can be seen that for a quarter inch wide tape, each band labeled 1, 2, 3 and 4 is 1/.16th inch wide but the actual recorded width is smaller. This prevents adjacent tracks from being heard on the tracks being played.  I think it's called cross-talk.

The quarter track head shown is a new one.

The head on the right is a half track.  Again, the blue and red marks are 1/8th inch wide for a quarter inch wide tape and the recorded bands are smaller for the same reason stated previously.

This is the same half track head shown above and has serious wear. Those dark spots on the upper track look like holes but this hasn't been confirmed. That condition wasn't realized until these photos were taken through a microscope
.

This head is obviously badly worn.  The uneven wear is due to the zenith being improperly set.  Zenith, sometimes called tilt, is the forward/backward angle (tilt) of the head face.  On a well designed machine, the head block is precision machined so this condition is minimized.  Also, the plate to which the heads are mounted is separate from the plate that fastens to the frame.  This upper plate has set screws to further correct zenith. Once set, they should not tampered with.  There are other screws, slotted pan head to adjust the height and azimuth.  This head seems to have come from a machine in which either the aforementioned set screws were adjusted or the heat height was improperly set, most likely the latter.
When replacing heads, the height may have to be corrected.  This can be done without affecting zenith if the front and rear screws are turned in the same direction and by the same amount.

A third slotted pan head screw is provided to adjust azimuth.  This is to set the gap in the head to be perpendicular to the longitudinal axis of th4e tape.

Theoretically, these two adjustments shouldn't affect each other but it's better to check.

This head (right) is very evenly worn and is at the point of needing replacement or relapping.
The casing above and below the part of the head that contacts the tape is machined below the surface of the head. This serves two purposes. 1. It may serve as a guide to when the head needs replacement or relapping, when the wear width is equal to the width of this flat surface. (left photo)

2. It prevents the edges of the tape from becoming damaged as the head wears. When this happens, a channel in the head surface will be the width of the tape and due to variations in tape width and tracking errors however small, the edges of this channel will damage the tape.

The next photo shows this phenomenon.

This head comes from a TEAC X-1000, of which I have two.  When fine black particles were seen on the microphone level control knob, it was found to be tape shedding from the playback head, which is a few inches above that knob.  These heads have no machined edges so the tape will wear a channel in the surface.  One of  those playback heads is at the right.  It's a quarter track.
The top edge that caused the tape damage is circled.  It's the dark circular segment above the yellow line.

All heads in both these machines were replaced.  I was lucky to find them at TEAC some 25 years after purchasing the first machine in 1984.  Capstan pinch rollers and drive belts were also purchased.

There is one more mechanical adjustment for the heads; it's called WRAP (tangency)  It's ensures the taps contacts the head evenly on each side of the gap.  When installing new heads, tighten the screw(s) holding the head to the upper plate just enough to stabilize the head but not enough to prevent its being rotated by hand.
Use a china marker to coat the head surface.  Play an old tape for about 10 minutes  and then inspect the wear mark in the marker that was made by the tape.  If it's the same distance on each side of the gap, you're all set to go.  Remove the head block (3 screws), tighten the head screw(s), clean the heads with alcohol and re-install the block.  Be sure to use the same tension on the screws through each change of the head block to the frame

If the side spacing is uneven, remark the head and rotate in the appropriate direction and repeat the above as many times as it takes to get the wear mark even about the gap.  Once achieved, carefully remove the whole block and tighten the screw(s).  This and head height is best done before soldering the wires to the heads.

I pluralized screw as some heads have two screws on the bottom and these may automatically set wrap.  I have not seen such heads but they are listed in the service manual parts list..

Now you're set to do azimuth, assuming the machine records, play back and runs.  If starting with a machine recently purchased, all functions should be repaired prior to replacing the heads.  Heads can be checked for assumed operability with an ohmmeter.  Remove wires HB1 thru HB7 from the left side of the VU meter board and remove both reproduce (playback) amplifiers.  This will isolate all the heads from the other circuits.  The wires attached to the heads can stay attached.  Erase = 0.6W Record = 15W Play = 96W (each channel and +/- 10%)

I did find one machine, a high speed half track with playback heads that measured 157W on each track and it works.  I was unable to find any info as to why.

1.  LMS Loudspeaker Management System
A circuit board measuring about 15" by 5" and fitting into an ISA computer slot. (No longer seen on mother boards)  It was purchased in 1994 at a cost of $1300.  Comes with a microphone (calibrated) by installing a file into the system once the software is installed.  The original ran under DOS but later, the Windows 98 update was purchased.  It has a myriad of routines.  The one most used was the determination of Thiele/Small loudspeaker parameters.  Another tool  is the gated response utility which will pulse the loudspeaker with tones from any bandwidth.  After giving the system the microphone distance from the speaker and the distance of the nearest reflection, the pulse it turned off before the reflection reaches the mic.  This simulates anechoic conditions.  Its limitation is when the wavelength becomes longer than the mic distance at which point the system can't get a good sample and eventually stops.  This happens around 300hz with the mic at a distance of 1 meter from the speaker.

It will also send a logarithmic sweep to an amplifier or tape recorder.  By placing then recorder in monitor mode, the recorded signal is picked up by the playback head and sent to the recorder's output and then to the LMS input.  This gives an actual response curve of the tape recorder.  I use it for setting bias and equalization adjustments.  No guesswork.

It will also calculate phase response of a speaker determined by data input and impedance response curves.  All calculations are done on the board so it's independent of processor speed.

An updated model ($3400) comes with 2 mics and has an input for a third along with two other inputs.  This allows the system to measure phase dirctly.  It is USB powered, hence portable.

Due to the passing of the owner/founder, the company has been shut down, including their website, email addy and phone.

木又 · 发表于 2023-9-22 08:48:02

打理完频响有这么好？

毕克 · 发表于 2023-9-22 10:29:08

Great!

烟雨夜 · 发表于 2023-9-25 18:13:18

牛逼，版主的英文也这么牛！我是看不懂，我就看了个寂寞

		自动登录	找回密码
密码			注册加入