Hi All
I assume someone here has done this already. I already sourced most of the trimmers for this job. I have already tackled the power supply too.
I am looking to re-cap these boards:
Switch Board
Input Amplifier
Record Amplifier (2x)
Oscillator
Record Relay
Playback and Drive Amplifier (2x)
Gee, this cap list is looking evil right now. Can anyone share their experience (and maybe their list) for this?
Some of the electrolytic values are hard to find too. Some of the 3,3uF values could probably be replaced by Film (MKP) once I look at the schematics. I am tempted to try Film caps in what looks like coupling positions. Any reason not to do this?
My machine is a high speed IEC one. Sounds really good. Fixing the meters was a challenge, but they are once again good.
I assume someone here has done this already. I already sourced most of the trimmers for this job. I have already tackled the power supply too.
I am looking to re-cap these boards:
Switch Board
Input Amplifier
Record Amplifier (2x)
Oscillator
Record Relay
Playback and Drive Amplifier (2x)
Gee, this cap list is looking evil right now. Can anyone share their experience (and maybe their list) for this?
Some of the electrolytic values are hard to find too. Some of the 3,3uF values could probably be replaced by Film (MKP) once I look at the schematics. I am tempted to try Film caps in what looks like coupling positions. Any reason not to do this?
My machine is a high speed IEC one. Sounds really good. Fixing the meters was a challenge, but they are once again good.
Hi Soulmerchant. The 3.3µF tantals are very hard to find but they can all be replaced by 4.7 µF electrolytics without any problem. As for replacing those caps with MKP, I have seen it done, but I don't see the point except spending more and taking up more space on the cards. Even being non-polarised capacitors does not add anything in these circuits which are designed in such a way that these caps will never see a negative voltage anyway. I would just replace them with good 4.7µ chemicals - in fact this is what I've done in all my Revox restorations, they measure and sound the same except maybe just a little more LF extension (not a bump), but practically unnoticeable in real life.
As for the other caps, usual advice applies: replace axial caps with axial caps whenever possible (essentially for practical and aesthetic reasons), replace with caps of the same voltage or more (the low voltages chemicals that were common in the '70s are hard to find now anyway), and don't replace a non-polarised capacitor unless you find it defective: they do not age, and the values of those located on the PB, REC and OSC boards are critical to set the time constants of the CCIR equalization and the tape bias.
You will probably want to replace the X2 capacitors on the relay PCB, especially if they are crackling. The replacement capacitors must be X2 type and nothing else! This is critical for fire safety.
As for the other caps, usual advice applies: replace axial caps with axial caps whenever possible (essentially for practical and aesthetic reasons), replace with caps of the same voltage or more (the low voltages chemicals that were common in the '70s are hard to find now anyway), and don't replace a non-polarised capacitor unless you find it defective: they do not age, and the values of those located on the PB, REC and OSC boards are critical to set the time constants of the CCIR equalization and the tape bias.
You will probably want to replace the X2 capacitors on the relay PCB, especially if they are crackling. The replacement capacitors must be X2 type and nothing else! This is critical for fire safety.
Kits are aviable in europe too.
No experience with this one but may be its worth a look
PREMIUM Reparatursatz REVOX A77 KOMPLETT Kondensatoren Trimmer Repairkit MK1-MK4 | eBay
No experience with this one but may be its worth a look
PREMIUM Reparatursatz REVOX A77 KOMPLETT Kondensatoren Trimmer Repairkit MK1-MK4 | eBay
IMO, ALL tantalums should be replaced. All audio path capacitors should be polypropylene wherever possible----above ~2.2uF, size becomes a real factor. For those larger values, use Nichicon ES bipolars---they have the lowest measured distortion of any electrolytic. For power supply (both smoothing and filtering), I recommend only Nichicon, Panasonic, or United ChemCon brands, and I try to stick to 105○c, 5000 hour caps.
Also check over at tapeheads.com---many users there have done these recap jobs.
Also check over at tapeheads.com---many users there have done these recap jobs.
Last edited:
Perhaps. But the engineers at Studer (for instance) thought it was important. And they some damn fine machines.
Studer, you said? Considering the number of electrolytics in the audio path of, let's say, a 990 console, I really don't think the minuscule THD difference was the reason why they used PP over PE. And they obviously did not choose to replace all the caps by PP - actually, if you take the caps in the audio path of a Studer as the criterion, you might easily conclude that 'lytics are the best audio caps in the world. Electronic engineering is not so simple, and there are many more factors to consider than you seem to believe.
Incidentally, one should not replace the film caps in a Revox unless you know exactly what you are doing, as these caps set important values (equalization, time constants, etc.) that are easy to ruin and difficult to readjust. The engineers who chose these caps were not bozos - in fact they were also the engineers who designed the Studers - and you are unlikely to know electronics better than them. To put it clearly: the supposed benefits of replacing caps by PP everywhere are not by far sufficient to take risks with a perfectly good Revox.
It is a good idea to replace the tantalums when restoring a Revox, not at all because they were bad caps but because they are now 50 years old and some of them may have failed. There is nothing wrong with tantalum caps, in fact they were better and more reliable than the 'lytics of the time. Revox (and Nagra, and Schlumberger, and Schoeps, as well as many other reputed manufacturers of TOTL equipment) knew how to avoid their biggest flaw, the crossover distortion.
Incidentally, one should not replace the film caps in a Revox unless you know exactly what you are doing, as these caps set important values (equalization, time constants, etc.) that are easy to ruin and difficult to readjust. The engineers who chose these caps were not bozos - in fact they were also the engineers who designed the Studers - and you are unlikely to know electronics better than them. To put it clearly: the supposed benefits of replacing caps by PP everywhere are not by far sufficient to take risks with a perfectly good Revox.
It is a good idea to replace the tantalums when restoring a Revox, not at all because they were bad caps but because they are now 50 years old and some of them may have failed. There is nothing wrong with tantalum caps, in fact they were better and more reliable than the 'lytics of the time. Revox (and Nagra, and Schlumberger, and Schoeps, as well as many other reputed manufacturers of TOTL equipment) knew how to avoid their biggest flaw, the crossover distortion.
That is a TOTALLY incorrect statement. Tantalum electrolytic capacitors are perhaps the most unreliable passive component ever put on a circuit board. They used them back in the day because they were cheaper and smaller in size for a given value than others. I don't think the engineers at that time had any idea of how bad they actually were, in the long run. They are notorious for their failure on power rails, shorting out and often causing much more damage to other components associated with their circuits. They have MUCH more distortion than than other types of caps (see the Bateman capacitor articles) when used in an audio path. DIY guru extraordinaire Rod Elliott thinks they are only good for the trash can. There is PLENTY wrong with tantalum caps. Their only saving grace is that they do impart a sort of distortion that some perceive as pleasant (Neve consoles, and some guitar amplifiers).
Last edited:
That was in the old days, with tape formulations that have been MASSIVELY improved upon since those years. Per the data sheet for RTM 900 tape :
https://www.recordingthemasters.com...00-technical-datasheet-IAQ-COM01-CLCOM-16.pdf
THD at nominal operating level (320nWb/m, 15 ips) is ~ 0.12%.
Might as well make the supporting circuitry as clean as possible, anyway. Otari engineers also thought to use only bipolar electrolytics in their audio path.
Last edited:
I do share your bad experience with Tantalum caps at that time. They were forbidden in my designs since the 80ies. My Revox A77 had loads of Roederstein electrolytic coupling caps, the ones in the cherry plastic cans. These dried out to ridicolous low capacitance and have been replaced by standard electrolytics. Replacing standard electrolytics - with their neglible distortions - by bulky foil caps makes as much sense as replacing copper speaker cables by silver foil imho.
I'll let you call Nagra and Schoeps to tell them that their designs are totally unreliable, and have been so since the 60's. I can't do it myself because then I wouldn't be able to explain them why all the stuff I have from them is still working perfectly. Maybe it's just because blanket statements like "tantalum caps are unreliable, period" are inaccurate, go figure.
A 3.3µF tantalum cap was not really smaller than a 3.3 µF electrolytic cap of the same years - I still have a few ones so I can compare. No, it's not a factor there.
Tantalum caps were much more expensive than chemicals even then, so cost was definitely not a a factor either. Example from a 1982 French ad: 10 µF, 25V tantalum caps cost 1.20 FF each, vs. 0.25 FF for chemicals. That's right, almost 5 times more expensive, and yet many engineers in their right mind used them, including for industrial, aerospace and military applications where sloppy engineering is not exactly welcome and tantalum caps were even more expensive. Aren't you curious to know why?
You are correct about one thing: the main failure mode of tantalum caps is short-circuit, it's in the specs. You are also correct that some engineers don't know what they are doing (I know a couple myself 🙄). But the vast majority, in spite of what you think, know a thing or two about reliability - it's part of the curriculum - and know the relationship between voltage, series resistance and MTBF - it's in the datasheets and application notes.
For instance tantalum caps in power rails are supposed to have voltage rating above the supply voltage by at least a given margin (stated in the technical sheets) to take into account the very low impedance, otherwise their life is shortened. Most of the knowledge about all this was already available at the time. But also, all too often, good designs get ruined by some glorified bean counters, cutting corners (and sacrificing reliability) to save a couple of pennies. I've seen it happen myself, and it explains a lot of early failures.
Most importantly, any machine is designed with an expected lifetime in mind (I'm not talking about built-in obsolescence, which is a completely different issue) and Revox did definitely not expect their A77 to be still in service 50 years later, so it is perfectly normal that you need to change some parts now. If they had designed it for 50 years, they would have designed it like a Nagra (also full of tantalum caps, just not the same type), it would have cost almost the price of a Nagra, and they would have sold much, much less A77's. So we should be grateful that they did not build it to last 50 years, which allows us to still have so many A77's and we only have to change a few caps, trimmers and ball bearings. Welcome to the world of real-life engineering!
I know well Bateman's paper, and truly enjoyed reading it. But you seem to have overlooked the fact that he did not really test the tantalum capacitors. He just tested an unnamed bead cap (whereas there are a bunch of tantalum cap technologies), in the worst conditions (no bias, plus one quick test with unknown bias giving unquoted values), and decided that he did not want to test tantalum caps any further. I respect his choice, but as a result this paper does not allow one to draw general conclusions about tantalum caps.
For you it's a matter of guru then, a religious thing. That is where our paths part: I am not at all interested in blindly following the opinion of one person, for me it's all about facts. I don't need a guru when I have the laws of physics (and psychoacoustics as far as audio is concerned), common and verifiable knowledge and experience to guide me. Maybe I'm just brainwashed by all these years of studying and practicing engineering.
Out of curiosity, how do you explain the sound, reliability and durability of, let's say, a Nagra IV-S and a Schoeps mic, both crammed with plenty wrong, only good for the trash can, tantalum caps?
You are absolutely correct that some pieces of equipment use these caps the wrong way - usually with reversed bias, or with no bias at all which is about the same. If you believe that those who design this stuff are clowns then you are wrong: on the contrary they know exactly what they are doing, deliberately, as a sound effect, just like they drive tubes way out of linearity for that lovely distortion. I don't believe this is the topic here. But I will grant you that the Revox had a reputation among some people to have a "certain sound", that they found pleasing and attributed to its tantalum capacitors. If there is any truth in this, the distortion measurements do not seem to show it - which is normal since the tantalum caps in the Revox are properly biased.
But the main question is whether, when restoring a Revox, you want to keep this particular character (if it exists) - which implies to stay with lowly tantalum beads - or if you want to switch to chemicals and thus make it neutral - on audiophile forums they say "lifeless and sterile". What is your priority? What should Soulmerchant do? I find it intriguing that after the fuss you made about the distortion of tantalum capacitors you don't show interest in this, as this is presumably the reason for the Revox being a Revox with its own character. This is probably the important question here, and a genuine restoration choice.
PS: for some knowledge about tantalum capacitors and their limitations, this link can be helpful: https://catalogs.avx.com/TantalumNiobium.pdf, pages 240-252. Knowledge in electronics and reliability is probably useful to read it. There are also interesting NASA reports on the reliability of tantalum caps, that you will easily find by searching "nasa tantalum reliability" on the 'net.
That is 20dB better than what I wrote, but still 40dB worse than the capacitors.
You might want to consider some other upgrades while you are recapping. I have an old REVOX A77 and had to replace the ice cube relays on the motor control board with relays that had 5 ampere contacts instead of 3 ampere contacts that would fail. Thats a simple plugin upgrade. Also some of the resistors on the motor control board were 5 watt and they would get really hot and burn the PC board. Replace them with 10 watt resistors and mount them off the circuit board so air can circulate around them. Lastly the friction material on the brake drum was a synthetic fabric and under hard braking would melt and seize the brake band. Replace the drum friction material with pure cotton that will not melt.