Hi Chris,
For the rest of transistor just try to get close Hfe for N/P pairs, nothing to critical.
I had small problem with my BC550/560 batch, could not get good matching, but that can be corrected changing current one of the input CCS. The CCS current setting resistor normaly is 820R, but I had to change positive CCS resistor to 1k to compensate for BC550/560 Vbe differences.
Damir
I'm using the RC combo - the cap is mounted on the solder side. HAve a look at the 2nd picture.
Using the diode variant the VAS current increases very fast above 100mA so I stopped using the BOM V1.1 variant and reverted to BOM V1.0.
BR, Toni
That increase was with thermaly coppled input supper pairs?
Can you tell what the VAS current is now?
BR Damir
Yes! It increases with increasing supply voltage very fast - above +/- 15V more as 100mA.
@ +/- 62V stabilized power supply:
VAS current is ~ 8.5mA
Bias 550mA total
Noise referred to 37.5V RMS / ~ 180W @8R
111dB unweighted/full bw
116dB 80kHz bW
117.5dB 20kHz bw
122dB A-weighted
IMHO really good values!
VAS suffers from exactly same sticking (@20kHz) as my enhanced VAS VFA (SA2013/SA2014) be4 using the VAS diode.
Next time I will add 2 BAV21 across the VAS transistors to recheck sticking behaviour.
BR Toni
Attachments
Last edited:
Hi Richard,
I described that in one of emails I have sent to you before. As I could not match BC transistors as good as it can be done, I increased CCS resistor(decreased CCS current). DC servo would correct it even with not changed resistor, but it has to work harder.
BR Damir
I am trying to make it all come together, I had to pull some resistors off the board because the supplier mislabeled the resistors. It was bad news all around. So I am building up a circuit to make sure it works before I put the transistors on the actual boards. I just hope it all works out for the best. I can change resistors and I can try different transistors with out breaking a sweat. I am learning a lot also.
Hi Toni,
Thank you for the information, I cant test that as I am out of my house for next 5 days.
I was afraid of that VAS thermal runaway when a diode was used instead VAS emitter resistor. With lower VAS current I could keep it quite stable, but did not try with higher current.
Ithink that a LED instead a diode and thermally coupled to the VAS transistor could do the trick, but that will nead sepparate higher power supply for the IPS.
By the way 8.5 mA is quite high for this VAS, mine first boards had it around 6 mA.
I dont like a diode across VAS, it increase distortion to much.
BR Damir
I am trying to make it all come together, I had to pull some resistors off the board because the supplier mislabeled the resistors. It was bad news all around. So I am building up a circuit to make sure it works before I put the transistors on the actual boards. I just hope it all works out for the best. I can change resistors and I can try different transistors with out breaking a sweat. I am learning a lot also.
I alwais check the resistors with an ohmmetter, just incase.
To have good desoldering tool is quite important for one hobbiest.
With a resistors is simple, transistors no so.
I have several desoldering tools, but the one that gave me the most trouble was a ground pad, that just sucked the heat away and left the hole full of solder.
Normally I check just before placement. This one slipped thru...I will get there SOON I hope.
THD+N versus power on 8 ohm load
All measurements done with:
Attached the raw THD+N data points.
(measurement #; Hz; db level of signal generator @600R; Vrms on output; THD+N in %)
BR, Toni
All measurements done with:
- +/-62V stabilized power supply.
- Bias ~450 to 500mA
- 8R resistive load
- VP7723-D (set to 80kHz bw range)
- THD+N plot
- test setup
- test setup
- test setup (20kHz measurement, ~ 170W@8R)
- 200kHz small signal square wave
- reference plot of VP7723-D audio analyzer
Attached the raw THD+N data points.
(measurement #; Hz; db level of signal generator @600R; Vrms on output; THD+N in %)
BR, Toni
Attachments
Last edited:
Hi Toni,
I like your test setup. I see that you soldered the OPS elcos from the back side, I suppose because the way the output transistors are fixed to the heat sink, to be away from the heat.
I can see that THD is higher at low output power, it is because how the analyser THD is higher at low output I suppose?
When you are going to do first listening?
BR Damir
Dear Damir,
the lower power values should be read "as good or better as". So values below 2 W output power would be better as shown in this plot.
The analyzer has its sweet spot from -12db to a maximum of +12db output. Today you will get a THD+N plot with lower output power where I use a 10db input divider to drive the analyzer at better/optimized output levels.
The capacitors are mounted on the back to be able to replace the output mosfets if damaged. If you someday decide to redesign the pcb may I place here some whishes:
About listening - I'm afraid I can't do that with highest quality because I drive my speakers with active Crossover - therefore I would need 4 amplifiers ...
the lower power values should be read "as good or better as". So values below 2 W output power would be better as shown in this plot.
The analyzer has its sweet spot from -12db to a maximum of +12db output. Today you will get a THD+N plot with lower output power where I use a 10db input divider to drive the analyzer at better/optimized output levels.
The capacitors are mounted on the back to be able to replace the output mosfets if damaged. If you someday decide to redesign the pcb may I place here some whishes:
- re-place the 8 electrolytics so it would be possible to remount the mosfets even if installed vertically. Add alternative 5mm footprint
- add some small heatsinks for the VAS
- place the big 470n MKP (RC-combo) to be able to remount the output drivers.
- add 7.5mm footprints for the 100nF caps
- add 2.54 mm footprints for the leds
- add 9mm footprint to be able to use MCP74 resistors
- move the gate resistors more closer to the mosfets
About listening - I'm afraid I can't do that with highest quality because I drive my speakers with active Crossover - therefore I would need 4 amplifiers ...
Dear Toni,
Some of your suggestion I already included in 1.2 rev.(never ordered PCB yet), like moved RC combo away from small heat sink. I dont have my PC with me and cant check for MCP7, but there is possibility to use wirewound source resistor(not sure if 9mm footprint). I prefare here to use 1% 2W metal film resitors (easy to order from China). I'll check for VAS heat sink.
BR Damir
THD+N versus power on 8 ohm load
... attached the lower power tests where the analyzer is better driven at its sweet spot.
For this test results compare with the VP7723D reference plot some post above. In the middle of the VP7723D plot starts the -12db and ends with +12db on the right side.
IMHO all very good values. Hopefully you will provide a redesigned pcb.
Have fun,
Toni
... attached the lower power tests where the analyzer is better driven at its sweet spot.
For this test results compare with the VP7723D reference plot some post above. In the middle of the VP7723D plot starts the -12db and ends with +12db on the right side.
IMHO all very good values. Hopefully you will provide a redesigned pcb
.Have fun,
Toni
Attachments
IMD and stability tests ...
IMD from left to right:
Stability checks:
All measurements taken with 8R resistive load.
BR, Toni
IMD from left to right:
- 0.0025% @ 10Vpp
- 0.0039% @ 20Vpp
- 0.0061% @ 40Vpp
- 0.0070% @ 80Vpp
- 0.0075% @ 102Vpp
Stability checks:
- 10kHz square rising edge (slew: 70V in 250ns ~ 280V/µs)
- 10kHz square falling edge (slew: 70V in 375ns ~ 200V/µs)
- 1kHz sinus mixed with square be4 clipping pos.
- 1kHz sinus mixed with square be4 clipping neg.
All measurements taken with 8R resistive load.
BR, Toni
Attachments
Dear Toni,
Thank you for your time spent on this very complex testing, I realy would like to have all that testing equipment.
You suggest to use different zobel, could you test with it?
BR Damir
You are welcome! Upcoming week maybe I have time to redo some tests with added zobel and/or VAS diode and some load tests using different capacitors.
BR, Toni
I am going back home today and I will check that. Can you tell what a power supply voltage you are going to use?