Sound Technology 1700B

[Trevor Wilson]

Well, it's been a few years. My 1700B has decided to cease operation, so
it really is time for a repair and full rebuild. Here's what I am planning:

* Replace all electrolytic caps in the unit. One of the main filter caps
is dead, so it's time to do the lot.
* Replace all the carbon composition resistors. A quick check on half a
dozen has shown a 15 ~ 25% upward drift. Why did ST use CC resistors?
I've been using cracked carbon since the 1960s.
* I was going to replace the OP amps with LME49710 OP amps, but, well,
you know. NLA. AD797s look like the best, available option. Any other
suggestions?
* Replace all the opto couplers.

[Dick Pierce]

On Sunday, January 2, 2022 at 3:51:27 PM UTC-5, Trevor Wilson wrote:
> Well, it's been a few years. My 1700B has decided to cease operation, so
> it really is time for a repair and full rebuild. Here's what I am planning:
>
> * Replace all electrolytic caps in the unit. One of the main filter caps
> is dead, so it's time to do the lot.

Probably not a bad idea.

> * Replace all the carbon composition resistors. A quick check on half a
> dozen has shown a 15 ~ 25% upward drift. Why did ST use CC resistors?
> I've been using cracked carbon since the 1960s.

Do remember that the ST1700B is a product of the late '60s and
early '70s to begin with. A LOT of these resistors are used as the
frequency-determining elements on the oscillator, so if they drift,
one result is frequencies are going to be off. If that's the only problem,
it's not a killer. I don't recall the eaxt topology but it may require
replacing two resistors per frequency.

And this is where you might run into problems. ST talked about how
components were carefully "hand selected" for optimum operation.
It may or may not be the case that those resistors that seem to have
drifte upwards by 15~20% may have been that way the day the
unit left the factory, so proceed with caution.

> * I was going to replace the OP amps with LME49710 OP amps, but, well,
> you know. NLA. AD797s look like the best, available option. Any other
> suggestions?

Unless they have measurable gone noisy or otherwise are known to be
bad, this is probably the LAST thing I would do, simply because the
amount of feedback around them tends to make differences less relevant.

> * Replace all the opto couplers.

I'd be VERY careful on this onje. Unless, as above, you can show they are a
problem AND if so, you can find suitable replacements (and their specs
are likely to be very specific, I'd leave them alone. Typically, in things like
the ST, they are used as part of the quadrature detectors and bridge balance
circuits, so unless they have similar transfer functions, newer, "better" ones
may, in fact, just not work.

Overall, my strategy would be to replace the obvious bad things first, put the
unit back together, test it thoroughly and then only proceed if there are still
issues.

Also, note that ST was very specific NOT to use vacuum desolderers (e.g.
"Soldapullit") on the unit because of their tendency to spray debris around
will-nilly. They recommend using capillary desoldering (solder wick or similar).

The ST1700B was, for its time, a truly qwonderful, almost miraculously good
unit. For about 5 years, I used one every day, measuring probably thousands
of amplifiers, preamps and so on. I had a lab where I built a very high-quality
passive RIAA pre-compensator so I could measure phone preamps with it,
I also had a switchable 50/75 uS preemphasis filter so I could use it with
my ST 1000 FM generator and do all sorts of measurements of FM tuners,
and so on.

A while ago, I acquired an Audio Sciences 32/44/48/88/96 kS/s 24 bit
professional broadcast sound card and got a licensed copy of the ARTA
measurement suite from artalabs.hr. It will do everything an ST1700B will
do, with THD residuals at least as low as the BEST 1700B, and will do
MUCH more (e.g., IHF IM, multi-tone IM, frequency response, FFT, RTA,
MLS, ...), and it will do it faster and more reliably. My ONLY complaint
about it is that it's Windows based, and Windows sucks.

Since then, I have been offered a couple of ST1700's and a ST1710,
one for small money, the other two free but for shipping. I turned
them down.

They were great units, 50 years ago...

+-----------------------------------+
Dick Pierce
Diverse Pursuits
Technical Engineering/Development
cartchunk.org
Boston - Spruce Head

[Trevor Wilson]

On 4/01/2022 6:17 am, Dick Pierce wrote:
> On Sunday, January 2, 2022 at 3:51:27 PM UTC-5, Trevor Wilson wrote:
>> Well, it's been a few years. My 1700B has decided to cease operation, so
>> it really is time for a repair and full rebuild. Here's what I am planning:
>>
>> * Replace all electrolytic caps in the unit. One of the main filter caps
>> is dead, so it's time to do the lot.
>
> Probably not a bad idea.

** I had intended doing this 20 years ago, but, you know, life.

>
>> * Replace all the carbon composition resistors. A quick check on half a
>> dozen has shown a 15 ~ 25% upward drift. Why did ST use CC resistors?
>> I've been using cracked carbon since the 1960s.
>
> Do remember that the ST1700B is a product of the late '60s and
> early '70s to begin with. A LOT of these resistors are used as the
> frequency-determining elements on the oscillator, so if they drift,
> one result is frequencies are going to be off. If that's the only problem,
> it's not a killer. I don't recall the eaxt topology but it may require
> replacing two resistors per frequency.

**All the critical resistors are either 1% or 0.25% tolerance
components. I assume they are metal film and I will not be touching
them. The less critical resistors are marked 5% tolerance and are carbon
comp types. Weirdly enough, some are from the E24 range and, of the half
dozen I tested, all have drifted by more than 15%. Even more weirdly, ST
shunted one carbon comp resistor with a second carbon comp resistor in
the power supply (723 regulator).

>
> And this is where you might run into problems. ST talked about how
> components were carefully "hand selected" for optimum operation.
> It may or may not be the case that those resistors that seem to have
> drifte upwards by 15~20% may have been that way the day the
> unit left the factory, so proceed with caution.

**Yeah, no. I don't buy that. As you know, carbon comp resistors are
notorious for drifting in value. In any case, I can see where close
tolerance resistors are used and they are clearly not carbon comp types.
I assume they're metal oxide or metal film ones.

>
>
>> * I was going to replace the OP amps with LME49710 OP amps, but, well,
>> you know. NLA. AD797s look like the best, available option. Any other
>> suggestions?
>
> Unless they have measurable gone noisy or otherwise are known to be
> bad, this is probably the LAST thing I would do, simply because the
> amount of feedback around them tends to make differences less relevant.

**It will be the last thing I do, simply because I will have to wait for
the OP amps and the Brown Dog adapters. John Curl reports a 10dB
improvement in noise, be simply switching to AD797 OP amps.

>
>> * Replace all the opto couplers.
>
> I'd be VERY careful on this onje. Unless, as above, you can show they are a
> problem AND if so, you can find suitable replacements (and their specs
> are likely to be very specific, I'd leave them alone. Typically, in things like
> the ST, they are used as part of the quadrature detectors and bridge balance
> circuits, so unless they have similar transfer functions, newer, "better" ones
> may, in fact, just not work.

**Yeah, sorry, I was unclear. On of the optos is faulty and must be
replaced. I figured on replacing the lot as a preventative measure, but
I feel you are most likely correct. If it ain't broke, don't fix it.

>
> Overall, my strategy would be to replace the obvious bad things first, put the
> unit back together, test it thoroughly and then only proceed if there are still
> issues.

**Certainly.

>
> Also, note that ST was very specific NOT to use vacuum desolderers (e.g.
> "Soldapullit") on the unit because of their tendency to spray debris around
> will-nilly. They recommend using capillary desoldering (solder wick or similar).

**I use a Hakko FR300 desoldering gun. It is an excellent device and has
never caused an issue with any product I've used it on. It will (and
already has) performed well with the ST1700B on the parts I have already
removed.

https://hakkousa.com/products/desoldering/fr-300-desoldering-tool.html

I prefer it to my old Hakkos, which used an outboard vacuum pump.

>
> The ST1700B was, for its time, a truly qwonderful, almost miraculously good
> unit. For about 5 years, I used one every day, measuring probably thousands
> of amplifiers, preamps and so on. I had a lab where I built a very high-quality
> passive RIAA pre-compensator so I could measure phone preamps with it,
> I also had a switchable 50/75 uS preemphasis filter so I could use it with
> my ST 1000 FM generator and do all sorts of measurements of FM tuners,
> and so on.

**Funnily enough, I acquired an ST1000A and 1100A last year for
AUD$100.00. I used one daily back in the 1970s. Great instrument.

>
> A while ago, I acquired an Audio Sciences 32/44/48/88/96 kS/s 24 bit
> professional broadcast sound card and got a licensed copy of the ARTA
> measurement suite from artalabs.hr. It will do everything an ST1700B will
> do, with THD residuals at least as low as the BEST 1700B, and will do
> MUCH more (e.g., IHF IM, multi-tone IM, frequency response, FFT, RTA,
> MLS, ...), and it will do it faster and more reliably. My ONLY complaint
> about it is that it's Windows based, and Windows sucks.

** 79 Euro! YIKES Almost an impulse (pun intended) purchase. Looks to be
very interesting. My LinearX LMS gave up the ghost several years ago and
I was considering the purchase of Clio Lite, but the ARTA software,
combined with a decent USB audio box should do the trick very nicely.
Thanks for the tip. That will be a no-brainer purchase.

Doesn't alter my view of a standard box type distortion analyser though.

I did buy a Quantasylum QA400 a few years back, but I managed to damage
it. I've been cautious about plonking more money into a QA402, but it
does look pretty decent.

>
> Since then, I have been offered a couple of ST1700's and a ST1710,
> one for small money, the other two free but for shipping. I turned
> them down.
>
> They were great units, 50 years ago...

**Yeah, that's true. They've certainly been bested by more modern,
computer based products. A mate has one of these though:

https://www.historyofrecording.com/Shiba_Soku_AD725D.html

Probably one of the best non-computer based test instruments ever built.

I want one.

[Trevor Wilson]

On 4/01/2022 6:17 am, Dick Pierce wrote:

>

> A while ago, I acquired an Audio Sciences 32/44/48/88/96 kS/s 24 bit
> professional broadcast sound card

**I can't seem to access their web site. Are they still in business?

[Dave Platt]

>> A while ago, I acquired an Audio Sciences 32/44/48/88/96 kS/s 24 bit
>> professional broadcast sound card
>
>**I can't seem to access their web site. Are they still in business?

http://www.audioscience.com/ works for me, first time!

[Trevor Wilson]

**Yep. That works. The weird thing is that it didn't work yesterday.

[Trevor Wilson]

On 4/01/2022 6:17 am, Dick Pierce wrote:

>

> A while ago, I acquired an Audio Sciences 32/44/48/88/96 kS/s 24 bit
> professional broadcast sound card and got a licensed copy of the ARTA
> measurement suite from artalabs.hr. It will do everything an ST1700B will
> do, with THD residuals at least as low as the BEST 1700B, and will do
> MUCH more (e.g., IHF IM, multi-tone IM, frequency response, FFT, RTA,
> MLS, ...), and it will do it faster and more reliably. My ONLY complaint
> about it is that it's Windows based, and Windows sucks.

**I've listened to you. The ST1700B is still in the works, but I have a
few other jobs to complete first. Based on your suggestions, I've
ordered one of these:

https://au.creative.com/p/sound-blaster/sound-blasterx-g6

It seems to provide very impressive performance, for not much money
(AUD$200.00). Independent tests show that it seems to meet it's
performance specs quite nicely:

https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-sound-blasterx-g6.7016/

As soon as the X-G6 arrives (probably Monday), I'll purchase the Arta
software. If all goes well, I may graduate to one of these in a couple
of months:

https://quantasylum.com/products/qa402-audio-analyzer

Again, independent reviews look (very) good:

https://www.audiosciencereview.com/forum/index.php?threads/review-and-measurements-of-quantasylum-qa401-audio-analyzer.8694/

I own a QA400, but it has been damaged and is limited in it's
capabilities. It has been retired.

Once more: Thanks for the tips and suggestions.