DIY Front End 2022

Best performance will be at high rails. But I personally would not push this one over maybe +/- 30V. The resistors (R7, R8) in the kit are meant to be used with the particular CCS J113's that they give you, in order to set the intended current through Q3 and Q4. A matched pair of K170's will work great at Q1 and Q2, use 'em if you got 'em.

edit Well maybe better stick to +/- 25V with the Toshiba's... I'm spoiled with the cascode in my FE where I don't have to lose sleep over the excess voltage thing... Pushing the drain to source voltage on the K170 much over 40V will likely break them.... "3 legged fuses"....
 
Last edited:
And dissipation starts to get out of hand for these little TO-92 devices when you start cranking these rails up in a project like this. The KSA992 and J113 may be OK voltage wise but the junction temps get pushed to a level that is probably not conducive to the long term reliablity of those guys. They have those little slip-on heatsinks but I'm not sure how effective they really are.
 
Hello frontend2022-builders,

I made a new 'start-up' with this nice circuit. Made some new pcbs with more possibilities for experiments.

At Q1/Q2 I used Toshiba 2SK170 GR with an Idss of 6,81mA. Very well matched.
I exchanged the KSA992 against a ZETEX FZT951. Running at 9.8 mA. Why? Because I wanted to test it... It is a very low noise BJT.
Some adjustments to the resistors to get the DC-offset before the outputcap C3 to 1-2 mV. Behaves stable.
Very important: railvoltages have been reduced to +- 16 V DC - I didn't want to cook those precious Toshibas :hot::redhot:
After the 47 Ohm resistors (R9 / R10) I was anywhere around +- 15.5 V DC.

Then I made some measurements with the Focusrite Scarlet and REW.
Tried inputvoltages 100 mV / 500 mV / 1v sinus at 1 kHz. At output of the frontend 2022 - a 10kOhm resistor.
With 100 mV sin in I measured slightly above 1 V sin out.
And I've got this - in the pics...
Perhaps I can listen to this little creatures this evening?...
And they need some more adjustments...
Only some infos about my journey into this circuit.
Have a nice sunday!
Dirk 🙂
 

Attachments

  • frontend2022_2SK170GR_FZT951_measurement.jpg
    frontend2022_2SK170GR_FZT951_measurement.jpg
    128.6 KB · Views: 441
  • frontend2022_2SK170GR_7mA_FZT951_2_100mVsin_in.jpg
    frontend2022_2SK170GR_7mA_FZT951_2_100mVsin_in.jpg
    83.5 KB · Views: 420
  • frontend2022_2SK170_7mA_FZT951_500mVsin_in.jpg
    frontend2022_2SK170_7mA_FZT951_500mVsin_in.jpg
    84.8 KB · Views: 324
  • frontend2022_2SK170GR_7mA-FZT951_1Vsin_in.jpg
    frontend2022_2SK170GR_7mA-FZT951_1Vsin_in.jpg
    84.9 KB · Views: 416
  • Like
Reactions: Monk55 and von Ah
Hello out there,

could adjust the DCoffset at output to below 1mV. They are hangin' on the better PSU.
Those 'babies' are playing music... :note::sax::eguitar:
Cheers
Dirk
 

Attachments

  • frontend2022_2SK170_FZT951_DCoffset1.jpg
    frontend2022_2SK170_FZT951_DCoffset1.jpg
    195.9 KB · Views: 287
  • frontend2022_2SK170_FZT951_DCoffset2.jpg
    frontend2022_2SK170_FZT951_DCoffset2.jpg
    165.3 KB · Views: 289
  • frontend2022_2SK170GR_FZT951_ playing2.jpg
    frontend2022_2SK170GR_FZT951_ playing2.jpg
    126.9 KB · Views: 296
Hello Ben Mah,

made a measurement with the other PSU. I think you are wondering about the pretty high noise floor above 1 kHz?
I have also a lot of cables / wires crosssing here...?
Cheers
Dirk
 

Attachments

  • frontend2022_2SK170GR_FZT951_100mV sin_in_betterPSU.jpg
    frontend2022_2SK170GR_FZT951_100mV sin_in_betterPSU.jpg
    85.8 KB · Views: 189
The latest measurement shows lower noise level, below the harmonic distortion. So the better PSU does measure better.

When I do FFT measurements, I average the results. That way the plot of the noise shows with a smaller range in levels, allowing the harmonics to be more visible if they are at approximately the same level as the noise.

I am still curious about the overall distortion levels. They appear to be quite high considering that 2SK170 were used in place of J113 at the input. The distortions are actually higher than what would be expected for J113. Could it be the Focusrite generator? You could do a loopback check of the Focusrite at its various output levels. Also the Focusrite XLR input is cleaner than the TRS input. Which input were you using?

And your 0.5V input plot showed higher distortion than the 1.0V input plot. Perhaps they were mislabeled?