Man, you are good. I just discovered this thread. This will make many people happy once you make it available for purchase or to make (if that are your plans). I wish you full success, and I am subscribed! As for the choice of system, anything from XP to 8 will do and will make us all happy. Personally I like 7. Please keep us posted!
Thanks for your encouraging words 
I am still working on the firmware and progress only in small steps as i have a lot of other work to do. I will keep you informed!
I am also interested to do some improvement to the Sys1 hardware. The Sys1 Generator is a masterpiece of engineering. The design is mutch more elaborate than the SG505 for example.
I believe that the multiplying DACs and the FET switches are the limiting factors in the design. The output transformer may also be a source of distortion, but this needs confirmation by measurements. I believe that it will not be easy to improve the generator by simply switching opamps. But a LM4562 or similar opamp with not too mutch current noise could improve things a little bit.
The analyzer part is a lot easier to improve. The input amplifier is a NE5534 and is determining the system noise and distortion.
Some other analyzers of that age (Shibsoku) use discrete FET input stages to lower noise.
A AD797 may be better in this place. It will improve noise and common mode distortion if the input attenuator is not used. If the input attenuator is used the high current noise of the AD797 will make noise worse, but distortion should still be better.
Unfortunately i am a lot of travelling at the moment and have not the time to try it out myself...
I am still working on the firmware and progress only in small steps as i have a lot of other work to do. I will keep you informed!
I am also interested to do some improvement to the Sys1 hardware. The Sys1 Generator is a masterpiece of engineering. The design is mutch more elaborate than the SG505 for example.
I believe that the multiplying DACs and the FET switches are the limiting factors in the design. The output transformer may also be a source of distortion, but this needs confirmation by measurements. I believe that it will not be easy to improve the generator by simply switching opamps. But a LM4562 or similar opamp with not too mutch current noise could improve things a little bit.
The analyzer part is a lot easier to improve. The input amplifier is a NE5534 and is determining the system noise and distortion.
Some other analyzers of that age (Shibsoku) use discrete FET input stages to lower noise.
A AD797 may be better in this place. It will improve noise and common mode distortion if the input attenuator is not used. If the input attenuator is used the high current noise of the AD797 will make noise worse, but distortion should still be better.
Unfortunately i am a lot of travelling at the moment and have not the time to try it out myself...
AP has done some pretty clever stuff with feedback through the output transformers, so don't casually dismiss them and assume that improvement will be easy or obvious.
I'd say that the same goes for the FET switches and MDACs - there are ways to run them so that their errors are actually quite minuscule, so again, it's not such a simple case of "these parts can stink" and that they actually _do_ stink in the complete circuit and implementation.
Still, don't let me discourage you from attempting to improve things. I'd try various simple little prototype sub-circuits before carving on an actual Sys1 or Sys2, just to see if the concepts hold up. Sure, small improvements like a more optimized noise floor or better basic linearity from modern ICs might be easy, but be careful to check for any unintended consequences from any improvements!
I'd say that the same goes for the FET switches and MDACs - there are ways to run them so that their errors are actually quite minuscule, so again, it's not such a simple case of "these parts can stink" and that they actually _do_ stink in the complete circuit and implementation.
Still, don't let me discourage you from attempting to improve things. I'd try various simple little prototype sub-circuits before carving on an actual Sys1 or Sys2, just to see if the concepts hold up. Sure, small improvements like a more optimized noise floor or better basic linearity from modern ICs might be easy, but be careful to check for any unintended consequences from any improvements!
If Udok succeeds with his USB APWIN cable, I would consider that the biggest improvement out of all. This is something no one has done and it will bring smile ear to ear of every AP Sys 1 owner. Almost all second market AP Sys1 units are sold without card, and eliminating need for it will be the second life for those units.
Yes, I worked at Jensen Transformers when the System One was a current product. I couldn't remember the specifics, so I emailed Bill Whitlock. He said I could post his response.
[. . .The post says “feedback around the transformer”, which is literally not true. Ap held a patent (now expired several years) that used feedback around the driving amplifier (not the transformer) to make the amplifier have a negative output impedance … just negative enough to effective drive the primary with true zero impedance. This has the effect of shorting out any non-linearity of the core material (which is modeled as a non-linear resistor in parallel with the primary winding) and makes the distortion of the transformer zero. In fact, the balance of the feedback to make the amplifier negative impedance is so delicate that resistance vs temperature of the primary winding must be accounted for and a special fine-wire winding with equal clockwise and counter-clockwise turns is included for the purpose. It’s an extremely clever design … someday I’ll do some app notes to implement the same principle on our standard output transformer line. . . .]
I'm lucky to just have gotten an old System One with a PCI card. It's not working, so I'll be digging around inside it in the future. I'm hoping it's just a power supply problem and easy to fix.
Love the idea of a USB adapter.
John
Interesting detail! Negative Zout seems to be an old concept to lower transformer distortion, probably back in the tube days someone already though about that. There also is a German patent from the late 70ies on that. From my own experiments I know that trimming the Zout is extremely critical and resistance drifts are a big problem when you really want to go lowest in distortion....
As far as i know AP is the only company who uses an output transformer for the generator.
HP 8903 uses a clever electronic balanced output circuit, which is described in a HP-journal.
The classical way is to use a separate floating power supply for the generator, if galvanic isolation or a true floating output is needed.
The capacitance between the primary and secondary winding of a power transformer can be made very low if the windings are separated and a shield is installed between the windings.
Does someone have some background information why AP chooses to use an output transformer? At first glance i see a lot of problems with the use of a nonlinear magnetic core in a low distortion analyzer.
Thx, Udo
HP 8903 uses a clever electronic balanced output circuit, which is described in a HP-journal.
The classical way is to use a separate floating power supply for the generator, if galvanic isolation or a true floating output is needed.
The capacitance between the primary and secondary winding of a power transformer can be made very low if the windings are separated and a shield is installed between the windings.
Does someone have some background information why AP chooses to use an output transformer? At first glance i see a lot of problems with the use of a nonlinear magnetic core in a low distortion analyzer.
Thx, Udo
AP's patent was to specifically address the non-linearity. Isolation through a transformer can be a really good solution and protect your internal circuitry as well. The distortion AP achieved with the transformer was state of the art for many years. The transormer was complicated and expensive. Jensen will build a transformer like that for you today, now that the patent has expired and then you need to build all the stuff around it. I tried but while it worked the product was too expensive and got scrapped.
Lundahl mentions a similar german patent DE2901567 from 1979 in the datasheet of the LL5402.
There the primary current is measured. By use of positive feedback the distortion is nulled.
It has the same problem with temperature and requires a separate winding for temp. compensation as the AP approach.
There the primary current is measured. By use of positive feedback the distortion is nulled.
It has the same problem with temperature and requires a separate winding for temp. compensation as the AP approach.
FWIW I'm dusting off a half-finished project on AP sys 1 graphing:
http://www.diyaudio.com/forums/equipment-tools/235883-anyone-know-parser-audio-precsion-system-one-files-new-post.html
System One Analyzer Noise
Hi Udok
The noise floor of the System One can be improved by changing the NE5534 with AD797. The compensation must be changed. I have only had one system that didn't have a noise floor improvement, typically 2 dB the best has been 2.7 dB.
FYI: The Generator output stage are the same for Sys1, Sys2, Sys2c & 2700.
I repair / service System One
Hi Udok
The noise floor of the System One can be improved by changing the NE5534 with AD797. The compensation must be changed. I have only had one system that didn't have a noise floor improvement, typically 2 dB the best has been 2.7 dB.
FYI: The Generator output stage are the same for Sys1, Sys2, Sys2c & 2700.
I repair / service System One
Udo,
Beautiful work so far!
Not being critical at all, but when I was looking at the pics of your board I noticed what *might* be a solder bridge. I figured I'd post you a pic of what I noticed just in case you happen to find something not working as planned when testing.
Again, AWESOME job!
Kind regards,
Steve
Beautiful work so far!
Not being critical at all, but when I was looking at the pics of your board I noticed what *might* be a solder bridge. I figured I'd post you a pic of what I noticed just in case you happen to find something not working as planned when testing.
Again, AWESOME job!
Kind regards,
Steve
Attachments
Autopoiesis ,I´d be interested to see more automation of S1´s output, and an USB interface too , so please do continue .
even at 25 years , the hardware can be kept working well for a long time,
with some careful maintenance.
A note to those who have or get hold of an old one:
Do change ALL the decoupling tantalum caps directly connected to/ across supply power - there are not very many of them, and they kill copper tracks if they blow.
To replace , good new longlife electrolytics will be fine, almost the same size as T´s these days. Those T´s found in high resistance series connections may just die quietly - if ever - they do last a long time. No need to change these unless dead.
I guess there is a lot of still workable but under-used S1 ´s
because of the lack of clever software and the USB interface to better use their capabilities.
HW specs are still close to the state of art.
AP naturally has little interest in these competing with demand for newer products.
even at 25 years , the hardware can be kept working well for a long time,
with some careful maintenance.
A note to those who have or get hold of an old one:
Do change ALL the decoupling tantalum caps directly connected to/ across supply power - there are not very many of them, and they kill copper tracks if they blow.
To replace , good new longlife electrolytics will be fine, almost the same size as T´s these days. Those T´s found in high resistance series connections may just die quietly - if ever - they do last a long time. No need to change these unless dead.
I guess there is a lot of still workable but under-used S1 ´s
because of the lack of clever software and the USB interface to better use their capabilities.
HW specs are still close to the state of art.
AP naturally has little interest in these competing with demand for newer products.
Just as an FYI. AP does offer an APIP-USB box ($350). I contacted AP a while back asking about compatibility with System Two analyzers. Here's their response:
AP also has a FAQ on the topic: AP software Windows compatibility
For the SYS 2400 and beyond, the AP USB interface works quite well. I've used it with the SYS-2700. However, for the older systems, you're hosed (unless Udo productizes his cable).
~Tom
AP also has a FAQ on the topic: AP software Windows compatibility
For the SYS 2400 and beyond, the AP USB interface works quite well. I've used it with the SYS-2700. However, for the older systems, you're hosed (unless Udo productizes his cable).
~Tom