The SG505 is an outstanding very low distortion state variable oscillator SVF developed by Bruce Hofer while he was at Tektronix before being a co-founder of AP. His use of the SVF inspired most of the circuitry in the THD analyzer I designed (Audio article available on my website). As described below in another post, JFET Q1620 amplifies and inverts and amplifies the error signal at the positive input to U1520. This enhances the net error signal applied to the op amp and further reduces the already low distortion of the 5534. The feedforward/feedback arrangement of R1521, R1423, pot R1510 and R1525 comprises what is often called a Baxandall volume control, in which the positive input of the op amp is usually grounded. The distortion-reduction circuit of Q1620 is just for further reducing distortion of the output buffer stage, and is not necessary if the already-low distortion of the 5534 is acceptable (good enough for more than 90% of applications). If not the right JFET is chosen, this circuit may not function correctly and actually increase distortion. You can disable this circuit and still get excellent distortion results. You can disable it by just removing the JFET from the circuit. See if doing this makes the circuit work properly. If you do so, you should probably short R1524 to keep the positive input of the op amp at a low impedance to ground, as normally done in the Baxandall volume control. Note that in general use, the Baxandall volume control provides a good approximation to a logarithmic gain vs. pot rotation using a simple linear pot over a very wide range of gain, making it excellent for use as an audio volume control.
Cheers,
Bob
Thank you Bob and Alex for the insights.
Now that we understand the purpose of the JFET, the next step is to find a replacement for the 151-1025-00 that is used. Unfortunately, I cannot find any details about it, nor for the original source, the SPF3036. It would greatly help if somebody would be kind/brave enough to measure the Gate voltage and the Drain voltage in the circuit or better yet, lift that JFET from any Tek instrument and put it on a Curve Tracer, or as a minimum, measure the Pinch off voltage and the rDS(on) for starters.
My goal is to see how far I can get the original SG505 working without making any changes at this moment, so I'm still trying to get the original output amplifier to work as close to the specifications as possible, before I start to use possible improvements, or use Bob's output amplifier section. 😉
So to continue with that quest, if I may in this thread, here is what my mouse-pal Bud (an ex-LT chip designer) found out while using LTspice for that particular circuit.
Many thanks for the collective insights, it proves the value of this forum.
Paul
Now that we understand the purpose of the JFET, the next step is to find a replacement for the 151-1025-00 that is used. Unfortunately, I cannot find any details about it, nor for the original source, the SPF3036. It would greatly help if somebody would be kind/brave enough to measure the Gate voltage and the Drain voltage in the circuit or better yet, lift that JFET from any Tek instrument and put it on a Curve Tracer, or as a minimum, measure the Pinch off voltage and the rDS(on) for starters.
My goal is to see how far I can get the original SG505 working without making any changes at this moment, so I'm still trying to get the original output amplifier to work as close to the specifications as possible, before I start to use possible improvements, or use Bob's output amplifier section. 😉
So to continue with that quest, if I may in this thread, here is what my mouse-pal Bud (an ex-LT chip designer) found out while using LTspice for that particular circuit.
Many thanks for the collective insights, it proves the value of this forum.
Paul
Hi Paul,
I may have missed something, but what happened to the original Tek JFET? Is it still alive? If so, check its IDSS value.
While it was in the circuit, did you check its operating point, such as its in-circuit drain voltage and thus voltage drop across its load resistor? There may also be a chance that the original part was a selected JFET. IDSS and pinchoff voltage can span a large range for JFETs. In a circuit like this, IDSS is probably more important than pinchoff voltage.
Don't rule out that there is something wrong other than the JFET.
Finally, to make sure there is not a problem of some other kind in the output buffer circuit, I would still disable the circuit by operating with no JFET and see if performance is good.
Make sure that the electrolytic coupling capacitor is the drain circuit is not defective in some way.
Cheers,
Bob
I may have missed something, but what happened to the original Tek JFET? Is it still alive? If so, check its IDSS value.
While it was in the circuit, did you check its operating point, such as its in-circuit drain voltage and thus voltage drop across its load resistor? There may also be a chance that the original part was a selected JFET. IDSS and pinchoff voltage can span a large range for JFETs. In a circuit like this, IDSS is probably more important than pinchoff voltage.
Don't rule out that there is something wrong other than the JFET.
Finally, to make sure there is not a problem of some other kind in the output buffer circuit, I would still disable the circuit by operating with no JFET and see if performance is good.
Make sure that the electrolytic coupling capacitor is the drain circuit is not defective in some way.
Cheers,
Bob
Hi Bob,
I do have two of the 151-1021-00 that is used in the SG505 AGC and found a few good alternatives.
The problem is that I do not have Q1620 p/n 151-1025-00 (the Motorola SPF3036) used in the output amp. Tek would have used another p/n if it was a selected version, so I'm very confident that it's "normal" part and it is also used in many other Tektronix instruments.
I tried several JFET's but they don't seem to improve the harmonic distortion, and the amp by itself produces significantly more distortion than the generator itself, that I found to be within the specifications. As I mentioned earlier, there's something fishy about my output amp contraption. Before I can zoom further in, I need to see if I can find a suitable alternative for the JFET, and Bud seemed to have found one, the LSK170C. Unfortunately, that part is not available to me in Europe, and going through Mouser or Digikey will set me back at least $30 for S&H alone.
That's why I'm searching for a good and available equivalent.
In the mean-time, I found a user that has an SG505 and was kind and brave enough to make some measurements in his unit.
He was not willing (I fully understand and appreciate that) to unsolder the part and measure the rDS(on) and Pinch-Off voltage.
Do we have enough data now to find other JFET's than the one Bud found?
Having said all that, before I can really zoom into the effects of the JFET, I need to improve my Twin-T notch filter setup (Dick Moore design) and significantly reduce the hum & noise. That's a side project that I'm working on right now.
I will put more information about that and the DIY build of the SG505 on my own Blog to not pollute this thread too much.
My understanding is that there could be some general merit in finding a good alternative for the 151-1025-00 that is used in many Tektronix instruments so that's the "excuse" for elaborating here. If you guys rather see a dedicated post, I'll start that.
Many thanks for the interest, I'm kinda overwhelmed by the quality responses!
I do have two of the 151-1021-00 that is used in the SG505 AGC and found a few good alternatives.
The problem is that I do not have Q1620 p/n 151-1025-00 (the Motorola SPF3036) used in the output amp. Tek would have used another p/n if it was a selected version, so I'm very confident that it's "normal" part and it is also used in many other Tektronix instruments.
I tried several JFET's but they don't seem to improve the harmonic distortion, and the amp by itself produces significantly more distortion than the generator itself, that I found to be within the specifications. As I mentioned earlier, there's something fishy about my output amp contraption. Before I can zoom further in, I need to see if I can find a suitable alternative for the JFET, and Bud seemed to have found one, the LSK170C. Unfortunately, that part is not available to me in Europe, and going through Mouser or Digikey will set me back at least $30 for S&H alone.
That's why I'm searching for a good and available equivalent.
In the mean-time, I found a user that has an SG505 and was kind and brave enough to make some measurements in his unit.
He was not willing (I fully understand and appreciate that) to unsolder the part and measure the rDS(on) and Pinch-Off voltage.
Do we have enough data now to find other JFET's than the one Bud found?
Having said all that, before I can really zoom into the effects of the JFET, I need to improve my Twin-T notch filter setup (Dick Moore design) and significantly reduce the hum & noise. That's a side project that I'm working on right now.
I will put more information about that and the DIY build of the SG505 on my own Blog to not pollute this thread too much.
My understanding is that there could be some general merit in finding a good alternative for the 151-1025-00 that is used in many Tektronix instruments so that's the "excuse" for elaborating here. If you guys rather see a dedicated post, I'll start that.
Many thanks for the interest, I'm kinda overwhelmed by the quality responses!
Q1620 is not rocket science. Given that R1620 is 510, the centered Vd is half the supply, 15/2=7.5 and Q1620 Vgs=0, the JFET is operating at Idss, so that the Idss should ideally be 7.7/510 ~=14.7mA. But if it's as little as 5mA or as much as 25mA, that's fine too, because the signal at the op-amp inputs is a "zero point", ie damn small. As long as Idss is not more than 25mA+ then you should be fine. So avoid J111, but J112 and J113 should be "OK". Note that 2N4416 Idss max is 15mA= perfect. You may want to find another JFET with Idss max = 15mA, or just hand select one. All it takes is a meter and bench power supply.