Thanks for that - when you say 150mA, I assume that is per transistor = 150mV across the 1R emitter resistors. That is 6.15W per output transistor. Works out to 98watts per channel dissipation into the heatsink at your rail voltage..
That doesn’t seem like enough power dissipation to cause the 4U by 500mm deep heatsinks to reach 50 C. I would have expected a total of 2 Amps or more needed to reach that point.
Per Gianluca's paper -
A 5U/500 @100W is ~20C above ambient under pretty much ideal conditions when measured on the fins
A 4U/400 @100W is ~30C above ambient under pretty much ideal conditions when measured on the fins
A 5U/500 @100W is ~24C above ambient under pretty much ideal conditions when measured on the hottest part of the sink
A 4U/400 @100W is ~36C above ambient under pretty much ideal conditions when measured on the hottest part of the sink'
(Note - As I've said in other threads, I can't see in the paper where the actual ambient temps were measured... I ballparked it from the graphs presented)
So depending on airflow, measurement technique, ambient temps... I don't see it being too far off if a 4U/500 would be somewhere in the middle of those two.
It also matches my anecdotal experience.
Edited to add - For the paper... only one heatsink was loaded... so in practice, temps may be slightly higher with both sinks loaded, since the temp inside the chassis would likely rise. Anyway... it's all just ballparks.
A 5U/500 @100W is ~20C above ambient under pretty much ideal conditions when measured on the fins
A 4U/400 @100W is ~30C above ambient under pretty much ideal conditions when measured on the fins
A 5U/500 @100W is ~24C above ambient under pretty much ideal conditions when measured on the hottest part of the sink
A 4U/400 @100W is ~36C above ambient under pretty much ideal conditions when measured on the hottest part of the sink'
(Note - As I've said in other threads, I can't see in the paper where the actual ambient temps were measured... I ballparked it from the graphs presented)
So depending on airflow, measurement technique, ambient temps... I don't see it being too far off if a 4U/500 would be somewhere in the middle of those two.
It also matches my anecdotal experience.
Edited to add - For the paper... only one heatsink was loaded... so in practice, temps may be slightly higher with both sinks loaded, since the temp inside the chassis would likely rise. Anyway... it's all just ballparks.
Interesting, I need to remember that I’m using less voltage, so my current figures are higher for the same wattage.
My F6 is dissipating about 93W per side in a smaller chassis.
My F6 is dissipating about 93W per side in a smaller chassis.
The stress test that I ran on my Stasis power supply used a resistive load that drew 2.2A from each 36V rail, one channel at a time. I don't plan to pull that much with the actual 8-deep channel boards. Maybe 200 mA per transistor; we will see.
Yea. 100 watts is a good happy medium of good power while staying in Class A. i built this for my father who wanted more wattage. If I had built this for myself I would have gone for 50 watts and more bias. Using Pico’s chassis id guess you could get close to 150 watts dissipation (per side) if the rail voltage was lowered or you used the 6 deep boards instead of the 8 deep. Also doesn’t help I live in Texas ha
You might consider getting a KL284 dual channel DC electronic load from eBay or elsewhere. Each channel can dissipate up to 200 watts (so 5.5 amps from +36V rail, plus another 5.5 amps from -36V rail). It's got temperature sensors on its heatsinks, which operate some BIG fans with lots of airflow. If I'm remembering correctly, it'll run the fans at four different speeds depending on temperature: 100% off, 33% on, 66% on, and 100% on. With a well thought out amount of hysteresis to prevent chuffing or yodeling among speeds.
Just dial up the current you want it to absorb from each channel (shown on the meters), hit the enable buttons, and presto, it does what you told it to do. While telling you the both the voltage and the current in each channel. So you can measure "sag" vs current if you wish.
I bought its little brother KL283 about ten years ago, and have absolutely loved its performance. OBTW the front panel meters have always been better than 1% accurate (!!) whenever I've double-checked them.
Just dial up the current you want it to absorb from each channel (shown on the meters), hit the enable buttons, and presto, it does what you told it to do. While telling you the both the voltage and the current in each channel. So you can measure "sag" vs current if you wish.
I bought its little brother KL283 about ten years ago, and have absolutely loved its performance. OBTW the front panel meters have always been better than 1% accurate (!!) whenever I've double-checked them.
Last edited:
That is a good price on the KL284. I have looked at DC electronic loads from Circuit Specialists and they were always kinda pricey, even when there was a sale. Kept me from considering one.
I have .380 mv across the emitter resistor and 40 output devices. Rails are +/- 45volts.
As Nelson said in a reply to my question back on page 1 of this thread with regards to max or recommended bias.
A quick request for expertise with BOM for ZM Stasis boards;
1. Is there anything special about C4 10uF, 63V on ZM's front end and does anyone have a suggested part?
2. any suggestion for P1 & P2 500 ohm trimpots? Bourn top adjust ... power rating?
3. any suggestion for SMD capacitors C1 & C2 for output stage
4. can someone propose a part for C3 100nF 100V for the output stage
Many thanks
1. Is there anything special about C4 10uF, 63V on ZM's front end and does anyone have a suggested part?
2. any suggestion for P1 & P2 500 ohm trimpots? Bourn top adjust ... power rating?
3. any suggestion for SMD capacitors C1 & C2 for output stage
4. can someone propose a part for C3 100nF 100V for the output stage
Many thanks
1. decent brand name, "just" filter for biasing voltage potential of cascodes for input LTP
2. plain vanilla multiturns, screw on top; you can even use either pins in line or pins in zigzag
3.1206 size, 100V
4. choose by form factor (size), decent brand name; nothing fancy needed
2. plain vanilla multiturns, screw on top; you can even use either pins in line or pins in zigzag
3.1206 size, 100V
4. choose by form factor (size), decent brand name; nothing fancy needed
1. is a simple electrolytic
2. 👍
3. I know nothing about SMD - is there such a thing as Polypropylene and if there is does it matter?- there seemed to be about a 1000 possibles for that!
4. OK thanks
ordered 400mm 4U and a bunch of 2SA1695 & 2SC4468...
Many thanks ZM
2. 👍
3. I know nothing about SMD - is there such a thing as Polypropylene and if there is does it matter?- there seemed to be about a 1000 possibles for that!
4. OK thanks
ordered 400mm 4U and a bunch of 2SA1695 & 2SC4468...
Many thanks ZM
all 4 above, regarding choice - just buy cheapest, taking in account criteria I explained in my previous post
My Chassis is home made using heat sinks that has 3" long fins. And the sinks run at 50 to 55C, winter/summer. There is a picture of the Amp somewhere back in the thread. The larger power supply that was necessary to handle the power level with very little ripple helps to mantain that level of bias. The sound is excellent with the high bias.
Sorry to be a pain...
A quick parts question... can anyone point me to a heatsink for Q5 & 7 for the Stasis Front end? ZM has them designated as KK1 and KK2 D01S 12mm x 12mm. 30mm long. I am not sure what D01S designates and have spent some time looking and come up with:
Advanced Thermal solutions
ATS-PCB1072
ATS-PCB1014
Any preference or proven solution?
A quick parts question... can anyone point me to a heatsink for Q5 & 7 for the Stasis Front end? ZM has them designated as KK1 and KK2 D01S 12mm x 12mm. 30mm long. I am not sure what D01S designates and have spent some time looking and come up with:
Advanced Thermal solutions
ATS-PCB1072
ATS-PCB1014
Any preference or proven solution?
I found the ATS-PCB1025 at Digikey and ordered a few to have on hand. They are an appropriate size for the TO-126 package.