Good job, Jama 
The fuzziness of the lines is mostly a consequence of the fact that the same data is line-plotted several times and each sweep up and back may be ever so slightly different. I control this to some extant by setting my timebase to capture just one waveform using a 2500 sample record size. Screenshots are fuzzy too for more or less the same reason.
If one wanted perfectly smooth lines for illustration purposes or something like that, the best way would be to plot them as point clouds and use a polynomial trend line in excel to represent a sort of average through the data. In this manner you could also extend the curves by projecting a few periods into the future.
But I digress...
The fuzziness of the lines is mostly a consequence of the fact that the same data is line-plotted several times and each sweep up and back may be ever so slightly different. I control this to some extant by setting my timebase to capture just one waveform using a 2500 sample record size. Screenshots are fuzzy too for more or less the same reason.
If one wanted perfectly smooth lines for illustration purposes or something like that, the best way would be to plot them as point clouds and use a polynomial trend line in excel to represent a sort of average through the data. In this manner you could also extend the curves by projecting a few periods into the future.
But I digress...
I did a test with the L'Amp circuit today, using the attached schematic (first one) with a 2SK182ES, and a simple regulator board for the bias.
It worked, well, to a certain extent. I noticed that if I increase the rail voltage pass a certain value (around 26-27V in my test), voltage drop across R2 will start to drift up -> less negative bias -> drain current increase -> voltage drop across R2 increase further -> even less negative bias -> ... until the current limiter of the power supply kicks in.
Does anyone encountered similar issue before, and what's the solution? (I guess the self biasing in the second schematic would probably avoid the issue, as increased drain current will increase the negative bias and forming sort of negative feedback, but I'm not very keen on another 30W of power dissipation by the biasing reststors).
It worked, well, to a certain extent. I noticed that if I increase the rail voltage pass a certain value (around 26-27V in my test), voltage drop across R2 will start to drift up -> less negative bias -> drain current increase -> voltage drop across R2 increase further -> even less negative bias -> ... until the current limiter of the power supply kicks in.
Does anyone encountered similar issue before, and what's the solution? (I guess the self biasing in the second schematic would probably avoid the issue, as increased drain current will increase the negative bias and forming sort of negative feedback, but I'm not very keen on another 30W of power dissipation by the biasing reststors).
Attachments
Hi Michael,
At 25V B+ supply when circuit is working nicely, voltage drop across 10k is around 0.1V (the DMM in the picture is showing the value). When B+ is increased to 27V, voltage drop increases to around 0.2V and the drift begins (I didn't record Vds, and reading at the power supply indicated Id of around 2A).
After a few minutes, the voltage drop across 10k could reach 1.5V and Id larger than 4A, at that point I shut down the power to avoid damaging the inductor.
I'll try replacing the device (maybe with a 2SK180, as I only have 2 pieces of 182ES).
Thanks and regards.
At 25V B+ supply when circuit is working nicely, voltage drop across 10k is around 0.1V (the DMM in the picture is showing the value). When B+ is increased to 27V, voltage drop increases to around 0.2V and the drift begins (I didn't record Vds, and reading at the power supply indicated Id of around 2A).
After a few minutes, the voltage drop across 10k could reach 1.5V and Id larger than 4A, at that point I shut down the power to avoid damaging the inductor.
I'll try replacing the device (maybe with a 2SK180, as I only have 2 pieces of 182ES).
Thanks and regards.
150uA is certainly a few times higher than most Tokins I've seen so far but not necessarily a show-stopper. It is temperature dependent and will probably come down some when your device and heatsink reach equilibrium, which is why sufficient heatsinking and good thermal contact between all the parts will be critical.
What you are seeing is the gate leakage rise as the part gets hot. It will move downward at some point as the heat is taken away by the heatsink. Take a look at Lab Hangout #4. You can also try a smaller resistor while you're figuring things out.
Set your B+ around the final value. Start with the bias all the way neg. and slowly adjust it upward over a few minutes while things heat up until you reach your desired current. Then observe it for an hour or so adjusting the bias to keep it on target. This experiment will clearly show you what's happening with the temperature dependence and give you an indication as to whether or not it might ultimately behave itself
What you are seeing is the gate leakage rise as the part gets hot. It will move downward at some point as the heat is taken away by the heatsink. Take a look at Lab Hangout #4. You can also try a smaller resistor while you're figuring things out.
Set your B+ around the final value. Start with the bias all the way neg. and slowly adjust it upward over a few minutes while things heat up until you reach your desired current. Then observe it for an hour or so adjusting the bias to keep it on target. This experiment will clearly show you what's happening with the temperature dependence and give you an indication as to whether or not it might ultimately behave itself
@ Jama Cwtim01 Michael
It's real joy to see all yours progress in Tokin L'Amp iterations
I have bring to my home today two huge heat sinks
his weight is 15 kg = 33 lb piece.
Body build and class A
Step by step i collect all components and mesure tools
for experiments with verticals. Nice hobby & relax.
Greetings
Christophe
It's real joy to see all yours progress in Tokin L'Amp iterations
I have bring to my home today two huge heat sinks
his weight is 15 kg = 33 lb piece.
Body build and class A
Step by step i collect all components and mesure tools
for experiments with verticals. Nice hobby & relax.
Greetings
Christophe
Hello, yesterday I measured leakage current of a thf51 during a time. I saw slowly rise, beginnig in 0.43uA and when heatsink reached its temperature (very hot) current was 2.4uA. In this moment readjusted pot, because Id down slightly (from 4.05A to 3.95A, more or less because I use analogic ammeter with 5A of full scale) and then seem drift was more slow. Vds was 25Vdc more or less. I use grid choke instead resistor (SAC Thailandia) with 7000H and 1K3 ohm DC. I will try with resistor again. But initially I used 10K and I did not saw Id drift. I will test it again.
Soundhappy your heatsink is very big, mine too, 50cm x 30cm x 8cm. But mine for 200wat of heat it is a bit small. I can not touch more of 3 seconds...
Soundhappy your heatsink is very big, mine too, 50cm x 30cm x 8cm. But mine for 200wat of heat it is a bit small. I can not touch more of 3 seconds...
I've used grid chokes with some (other) parts with more problematic gate leakage and they work nicely.
It's important to note that things "drifting" moderately while the amp "warms up" is all perfectly normal and should stabilize after an hour or two with adequate heatsinking and thermal contact. This sort of thing applies to nearly every simple Class A amp you will build unless, of course, the parts themselves are indeed bad and drift wildly.
When in doubt, suspect your heatsinking and thermal connections first, then your transistors.
It's important to note that things "drifting" moderately while the amp "warms up" is all perfectly normal and should stabilize after an hour or two with adequate heatsinking and thermal contact. This sort of thing applies to nearly every simple Class A amp you will build unless, of course, the parts themselves are indeed bad and drift wildly.
When in doubt, suspect your heatsinking and thermal connections first, then your transistors.
Last edited:
Well noted. The nice thing about this forum is that everytime I post a question, I'm learning something
I think the heat sinking in my test was pretty bad. The thermal pad was a very cheap one (5 USD for a roll of 50 feet) with mediocre thermal conductivity, there wasn't any thermal grease, neither at the pad nor between the L shape and the main heatsink, which was probably too small for the task anyway
I'll probably need to do some shopping on thermal stuffs before my next test
Ready to insert in the l'Amp amp the Tokin 2SK180 , here it is what it looks like .
I use 2 light bulbs in parallel each 500W 230V around 22 ohm or so , or probably a total of 3 light bulbs for 1150 W .
Edit : rail voltage = + 44 Volt , no degeneration resistor , 10k to the negative voltage . This particular piece sets rather well at vgs = -2,7 V .
I use 2 light bulbs in parallel each 500W 230V around 22 ohm or so , or probably a total of 3 light bulbs for 1150 W .
Edit : rail voltage = + 44 Volt , no degeneration resistor , 10k to the negative voltage . This particular piece sets rather well at vgs = -2,7 V .
Attachments
Last edited:
I ordered some mica and silicon compound from eBay, but is not patient enough to wait, so I put some CPU thermal grease on the 182ES and put it directly on the L-profile (without any insulator pad), tried again ... and the circuit behaved well now .
Some measurements:
B+ : 38V
Vds : 35.6V
Id : 1.95A
Bias : -7.45V
Leakage current : 8-10 uA
Transistor case temperature : 56 deg C (measured by a LM35 IC sensor)
Heatsink temperature : 53 deg C
Ambient temperature : 24 deg C (air conditioned)
Attached is the output waveform at 10W (yellow is input, blue is output), while I think there is some distortion visible, this waveform still looks quite nice to me.
Next step is to reinstall my notebook, run QA400 (distortion analyzer) on it and search for the sweet spot, but I will need to find a bigger heatsink to run Id higher.
.Some measurements:
B+ : 38V
Vds : 35.6V
Id : 1.95A
Bias : -7.45V
Leakage current : 8-10 uA
Transistor case temperature : 56 deg C (measured by a LM35 IC sensor)
Heatsink temperature : 53 deg C
Ambient temperature : 24 deg C (air conditioned)
Attached is the output waveform at 10W (yellow is input, blue is output), while I think there is some distortion visible, this waveform still looks quite nice to me
.Next step is to reinstall my notebook, run QA400 (distortion analyzer) on it and search for the sweet spot, but I will need to find a bigger heatsink to run Id higher.
Attachments
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.