Revox A700 repro IC

[TIMTAPE]

I've noticed my A700's playback noise is a little higher than my A77 and B77.
Unlike their discrete transistor design it uses an IC, TBA 931 - 3. 14 pin DIL. The PCB has + and - 12v rails.

I couldnt find much info about it on the www. Does anybody know if there is a replacement IC? Perhaps a more modern and lower noise type? Not sure how common is the 14 pin.

Any help appreciated.

Tim

[Michael Maurice]

Hi Tim,

I found this, it may be of help to you.

http://cgi.ebay.co.uk/A700-Replaceme...19114001r35401

Michael

[Michael Maurice]

A search on SN76131 brought this from Cricklewood electronics:

http://www.cricklewoodelectronics.co...=search&page=1

Food for thought possibly?

[Top Cap]

Looks like the LM1303N, TBA231,MC1303P or UA739 will also fit in.
A very informative pdf on the UA739 available here-

http://www.datasheetarchive.com/search.php?q=UA739

Les

[TIMTAPE]

Thanks both of you for your help. I'm not sure if the existing IC is noisy or not. It just seems to be (subjectively) re my other comparable machines, which use discrete trannies.

I thought it might be a good chance to upgrade the IC to a more modern and low noise variety, even if it means some modifications. I was thinking of maybe a NE5534AN or a TL072 or a better one even that those.

Cheers Tim

[SirenCloak]

Hi Top Cap, I work on tape machines myself. I have been searching for the Studer op-amp TBA931 used in the A-80's. I am trying to get one fixed for a friend and have identified several of these chips being bad. I am trying to find 72 of them to do a blanket replacement or a comprable part. I found a thread on here where you commented on some other chips you thought would fit in. Do you know for a fact that those chips will work as direct replacements? One was a Fairchild UA791 I believe.

Thanks so much for your help!

Bill

[Top Cap]

Not having worked on Studer machines or having any schematics, you would have to compare pin outs etc to see if there was a chance of fitting any later devices. You could fit an IC socket then compare different types. Technology has advanced and the chance of using a lower noise type may be possible.

[Ted Kendall]

If it ain't broke, don't fix it. I tried this decades ago with a 5532 on Studer's own substitution board for the 931 and got head magnetisation on switch-off for my pains. This eventually resulted in a Revox technical bulletin to the effect that their substitution board could be used anywhere in the machine but the rep amp and to swap 931s from elsewhere if trouble was encountered here.

[TIMTAPE]

Ted,

I'm not sure that "if it aint broke dont fix it" applies here. I started this thread in 2009 because the repro 931 in my A700 seemed noisy compared to an A77 and a B77 .

Both A77 and B77 use discrete repro circuits which is interesting because the B77 is a later model than the A700. The B77 uses IC's for other logic functions but not the repro amp or other audio functions on the cards.

Why would Revox revert to a discrete design over an IC one (whether the IC had been a 931 or not) unless there was some perceived problem with it in the repro amp?

I know the discrete repro amp design for the B77 creates a very slight DC current in the playback head but according to Revox this is so small as to be negligible.

It seems to me the problems with the substitution board mod in the repro circuit are a separate issue and dont necessarily mean the TDA 931 has a clean bill of health in that application.

I suspect that a reworking of the repro circuit, not just the substitution board, would fix the extra problem you mentioned, allowing modern IC's to be used. This would also address any S/N shorcomings which I mentioned.

Noise in professional reel to reel machines running at pro speeds is hardly an issue in their repro amps as there is plenty of play head signal for them to work with. Knowing that, I was surprised the 931 seemed so noisy in my A700. What would that same 931 sound like in a cassette deck application? One shudders to think.

Cheers Tim

[Michael Maurice]

I wonder Tim if the difference in the designs could have originated from different divisions within Studer.

The A700 is a semi pro machine based on the Studer pro machines and probably designed and built in Switzerland where as the A77 and B77 machines were Domestic machines probably built in Germany.

[TIMTAPE]

Interesting point Michael. Interesting too that the more consumer oriented A77 and B77, at least my examples of them, had lower noise than the more 'pro' A700 and other Studer designs.

[Michael Maurice]

I suppose if one of us was minded, we could redesign the playback amp but remember that the board has both ch1 and ch2 playback amps and eq for 3 speeds. Though with modern SMD transistors it could possibly be done.

[Ted Kendall]

The problem with the 5532 was that the asymmetrical collapse of power rails on switch-off sent a large current pulse through the head, magnetising it. This does not occur with the 931, presumably because the bias currents are much smaller.

The A67, developed in parallel with the A700, has a similar rep amp design. The machines were slightly in advance of the available technology but, given careful selection of 931, could be made to work perfectly well unless you needed to apply large amounts of top lift to retrieve an ancient tape.

The B67 substituted a three transistor discrete first stage, feeding an IC EQ stage not unlike that in the A700. Placing one of these in front of the existing circuit and reducing the gain around the 931 would serve two purposes -the first gain stage would be quieter, to the benefit of S/N ratio; and if the resulting improvement were not sufficient, the removal of the IC from the head circuit would permit substitution of a quieter IC.

Masochists could consider the use of the A810 rep preamp built into the block - this would bring similar benefits and probably reduce hum pickup, another A700 bugbear, for which the deployment of B67 shielding plates was recommended.

[TIMTAPE]

Ted,

I feel you are being very generous to Studer and Revox on this matter of noise in the 931 when used as a repro amp.

If the 931 was below par when used as a repro amp in this application surely it shouldnt have been used. Especially so when perfectly acceptable results had been obtained using standard discrete bipolar transistor circuits for years before that and for years afterwards.

Of course nowadays, low noise IC preamps are much better and I'm sure any problems of head magnetisation would be easily addressed by good design.

It seems the 931 was acceptable in its other roles in the Revox and Studer machines but its performance as a playback head amp was borderline to say the least.

I do appreciate the information you've supplied because it confirms my subjective impression of higher than normal head amp noise in my A700 wasnt just my imagining!

Cheers Tim

[Ted Kendall]

Generosity doesn't come into it - the machine was designed around the available or projected technology and was a quantum leap in value for money and complexity at the time. The use of ICs was central to keeping the cost and complexity within bounds. 931s came in three grades, according to noise. I take it you are using 931-3s in the rep amp? If not, swap one in from a less critical position.

As to modern ICs being better,the 5532 is still the class of the field for low noise audio work, so the problems of bias current remain. With the benefit of hindsight, the 931 wasn't as low noise in this application as it was intended to be. Bipolar discretes would have been a better solution at the head, as Studer themselves recognised, but the difference is that between a perfectly adequate configuration and one that is near-perfect. If nobody tried to innovate, we'd all still be running around in woad, as my old boss used to say...

If you must fiddle, note that the magnetisation problem with 5532s occurred on switch-off. The preamp worked perfectly well the first time it was used. The +and -12V rails decayed at different speeds, as the +5V was derived from the +12V, if memory serves, so it collapsed more quickly, causing a current pulse through the head. Some jiggery-pokery with diodes, downstream decoupling and defined decay through drain resistors may fix this, or it may not.

[TIMTAPE]

Ted,

Thanks for the extra information re fitting a 5532 though now that I know it's not as straightforward as it might have been (due to possible head magnetisation) I'll think twice before proceeding. I have other more pressing jobs!

Still since my A700 is otherwise the best 1/4" machine I have, for it to have a sub standard play head front end is a shame. These days I would only use it for capture to digits anyway where the playback chain is all important. Especially so since tape stock only got lower in bias noise as the years rolled on, making the preamp noise stand out more.

You say that on the B67 Studer reverted to a discrete front end, driving a 931, surely an admission on their part that even the lowest noise 931 wasnt up to the mark.

That to me was progress. Sometimes progress starts with admitting one has made a mistake and that there is a better solution, sometimes even a tried and true one. If the 931 was "near perfect" as first stage in the head amp, why did Studer later go to the trouble to place a discrete circuit in front of it?

Whichever way you slice it...

Tim

[Michael Maurice]

A manufacturer will design a piece of equipment based on research in a laboratory, he will then test his design and when he thinks he's got it right will start producing the item in bulk.

But after you've made a few thousand, certain problems that didn't show up in pre-production now do, so the manufacturer tries to find a solution and when he's found that solution incorporates it into the production line.

That applies to every product be it electronic or mechanical. Just look at the number of cars that have to be recalled for safety checks.

As Ted has said, Studer used the very latest technology available to them, dont forget the A700 came out in 1973 when IC's and especially large scale IC's were in their infancy.

Having found a noise problem, Studer went on to modify it in the next generation of recorders but would it have made sense to redesign the repro amp for the few that might have noticed the increased noise?

The other point to make is that theA700 uses +/-12V rails where as the A77 and B77 use a single 21V rail.

If as Ted suggests the problem with magnetisation occurs because of the different decays of the supply rail, you could always try doubling or tripling the 12V supply reservoir capacitor.

[Ted Kendall]

Exactly - and I didn't say that the B67 discrete front end drives a 931 - it drove an IC configuration not unlike the A700 rep amp, but the chip in question was the 5534. The 931 was not a particularly happy chip as it turned out, hence Revox's recommendation that it be replaced by a 5532 if it failed. However, the A700 rep amp and one location in the A720 tuner-preamp were not recommended for this mod, the suggestion being that functioning 931s be found elsewhere in the equipment and the substitute IC used in one of the non-critical positions.

[Michael Maurice]

After some discussion the moderators have decided that this thread should be closed.

There has been some aggression between two members which have resulted in two posts being deleted. This cannot be tolerated.