Look,
if you followed my build status you see, that i installed Kapton insulator washers between the Mosfet´s and the heatsink,
with an extreme low terminal resistance of 0,07K/W. This config. was untouched all the time.
With the addition of the copper blocks (with "Arctic Silver 5" thermal conductance paste in between) witch conduct the heat
from the sinks to the aluminum bottom, i lower the overall Rth.
That´s all.
cascaded regulators are an option.
Bleeders increase PSU ripple and hardly change output voltage. They give you nothing back for the heat they generate.
Relay switched bleeders that conduct after the mains' power is turned off could be of some use.
If input power is high and output power is low then the difference is made up of losses. High Losses = High Heat.
It does not matter whether the losses are through inductors, or resistors, or transistors or a combination. They all give out heat that must equate to losses.
Last edited:
Heat sink Insulation
Would someone please explain why the Mosfets need insulated from the Heat Sink or chassis?
If insulation is required, what insulating material is best to insulate and pass the heat to the sink?
Do the Diodes need insulated also?
Would someone please recommend part numbers for the insulators?
I don't know Laterals from Verticals
Would someone please explain why the Mosfets need insulated from the Heat Sink or chassis?
If insulation is required, what insulating material is best to insulate and pass the heat to the sink?
Do the Diodes need insulated also?
Would someone please recommend part numbers for the insulators?
I don't know Laterals from Verticals
Last edited:
The middle pin of power transistors is electrically identical to the backplate of the device.
Laterals have the backplate as the Source.
Verticals have the backplate as the Drain.
BJTs have the backplate as the Collector.
There are many thermally conductive and electrically insulating materials, of various prices, various thermal conductances and easy/difficult to handle.
Laterals have the backplate as the Source.
Verticals have the backplate as the Drain.
BJTs have the backplate as the Collector.
There are many thermally conductive and electrically insulating materials, of various prices, various thermal conductances and easy/difficult to handle.
Andrew, I think he understands very well. Perhaps you are taking his words too literally when he says that the sinks are cold. Such terms are always relative. Relative to before he added the copper with thermal paste, it is now much "cooler" or "less warm" perhaps you would have preferred he say.
He has now got some quite massive and well connected heatsinks on his mosfets, and at the present 200mA draw, it is surely pulling much less current than it could.
I do understand what you're saying - it's clear, and I am sure that Oliver understands the principle of what you are saying, that cold heatsinks AND a hot mosfet, means the sinks aren't connected properly, but it is unlikely that the mosfets are effectively unsinked and burning hot, because Oliver is very careful, and I am sure would connect them to the sinks very tightly, and as he has said already, he is using quality insulating pads and thermal paste.
Perhaps the problem is that he said they were "cold", which can mean "stone cold" or else "relatively cooler than before" which I took it to mean.
Oliver is German, don't forget. In Oliver's own language, he would normally have more specific technical terms to chose from.
He has now got some quite massive and well connected heatsinks on his mosfets, and at the present 200mA draw, it is surely pulling much less current than it could.
I do understand what you're saying - it's clear, and I am sure that Oliver understands the principle of what you are saying, that cold heatsinks AND a hot mosfet, means the sinks aren't connected properly, but it is unlikely that the mosfets are effectively unsinked and burning hot, because Oliver is very careful, and I am sure would connect them to the sinks very tightly, and as he has said already, he is using quality insulating pads and thermal paste.
Perhaps the problem is that he said they were "cold", which can mean "stone cold" or else "relatively cooler than before" which I took it to mean.
Oliver is German, don't forget. In Oliver's own language, he would normally have more specific technical terms to chose from.
Thanks Andrew. That's what I thought. Such losses will generate in the mosfets, right? I am concerned that I may need bigger sinks or lower current to compensate for my high voltages.
So bleeder resistors are of no use then, other than to drain the caps at power off, in your experience? Draining the caps is hardly necessary either, is it? I may leave them out then.
Hi Gary,
The mosfets need insulating only if they are sharing a heatsink or two. If they are 4x independent sinks, no insulation is required or desirable. If sharing a sink, they need electrically insulating from each other, in such a way that the heat is still conducted. Oliver tends to buy the best, so copy him - Kapton insulator washers.
Diodes need no insulation from 2x56R up to and including a 10R current resistor, which is 99% of us.
So, have been reading all the thread, and I would like your opinion on the following matters. I think I will try the 10R and the 3R3 versions:
Power transformer: I live now in Japan that has 100 V, but I will move sooner or later back to Europe. So I was thinking to buy a transformer 115/230 primary and 18-0-18 out. This will give my 15.6 V in Japan and 17 V in Europe. The alternative is to have a 15-0-15 V secondary and run at 13 V while in Japan. I can change resistors when I move.
I have to choose between the R26-22 and the R26-34 by DIYclub
What do you think is better ?
Potentiometer: I think I will choose a 20 K pot. Does it make sense to use a stepped attenuator ? advantages and disadvantages ? I'd love to install something that I can remotely control in the final build, suggestion ? Lightspeed is also an option.
I will use the buffer mainly between the DAC and the Power amplifier. The DAC (DAC-END2) has a tube output with 1 5687 with the sections in parallel and 1 mF output cap. I have various amp, but they all start with a tube (5842, 6N1P, 12AX7). I plan to play with i/v stages anyway.
Does it make sense to use two buffers before and after the pot. It looks like in the Pass Lab threads they think this is the best solution.
For the resistors I think I will use Caddok for the signal. I am tempted by the Foil one for the 220R, but they cost 15 $ any cheaper source ?
Also, I did not find a source for the pot. Any suggestion ? I find only 10k or 50k.
Output caps: Am I right thinking that with OPTs before the speakers there is no way of getting DC on the speakers ? Worst case scenario, if there is a problem with the buffer, the bias of the input tubes will change if I have some DC, although with leds on the cathodes this should not happen.
Thanks,
Davide
Power transformer: I live now in Japan that has 100 V, but I will move sooner or later back to Europe. So I was thinking to buy a transformer 115/230 primary and 18-0-18 out. This will give my 15.6 V in Japan and 17 V in Europe. The alternative is to have a 15-0-15 V secondary and run at 13 V while in Japan. I can change resistors when I move.
I have to choose between the R26-22 and the R26-34 by DIYclub
What do you think is better ?
Potentiometer: I think I will choose a 20 K pot. Does it make sense to use a stepped attenuator ? advantages and disadvantages ? I'd love to install something that I can remotely control in the final build, suggestion ? Lightspeed is also an option.
I will use the buffer mainly between the DAC and the Power amplifier. The DAC (DAC-END2) has a tube output with 1 5687 with the sections in parallel and 1 mF output cap. I have various amp, but they all start with a tube (5842, 6N1P, 12AX7). I plan to play with i/v stages anyway.
Does it make sense to use two buffers before and after the pot. It looks like in the Pass Lab threads they think this is the best solution.
For the resistors I think I will use Caddok for the signal. I am tempted by the Foil one for the 220R, but they cost 15 $ any cheaper source ?
Also, I did not find a source for the pot. Any suggestion ? I find only 10k or 50k.
Output caps: Am I right thinking that with OPTs before the speakers there is no way of getting DC on the speakers ? Worst case scenario, if there is a problem with the buffer, the bias of the input tubes will change if I have some DC, although with leds on the cathodes this should not happen.
Thanks,
Davide
Could this fit as pot choice ?
Hi-End Volume Control DIY Kit,LED, Remote Control /V-03 - eBay (item 290485343961 end time Nov-07-10 04:10:12 PST)
It says max out impedance 22K.
Thanks,
Davide
Hi-End Volume Control DIY Kit,LED, Remote Control /V-03 - eBay (item 290485343961 end time Nov-07-10 04:10:12 PST)
It says max out impedance 22K.
Thanks,
Davide
Wow! not cheap, those!
You could also try the highly regarded Vishay Naked Z foils if you're spending:
Vishay VAR-Series "naked" Z201 Z-FOIL RESISTOR
Regarding values, Nelson Pass originally had them at 1K, and they were reduced by popular consent/dissent, (based on reports of tighter bass with more drive) to 220R, which Papa agreed with I believe, warning however, that the lower value you go, you increase risk of parasitic oscillation. So, you can choose any similar value and be safe, and go up to 1K, but I can confirm that it sounds better near to 220R.
You could also try the highly regarded Vishay Naked Z foils if you're spending:
Vishay VAR-Series "naked" Z201 Z-FOIL RESISTOR
Regarding values, Nelson Pass originally had them at 1K, and they were reduced by popular consent/dissent, (based on reports of tighter bass with more drive) to 220R, which Papa agreed with I believe, warning however, that the lower value you go, you increase risk of parasitic oscillation. So, you can choose any similar value and be safe, and go up to 1K, but I can confirm that it sounds better near to 220R.
What about using one or two buffers ? (before and after)
Lucas, I saw that, I was looking for a cheaper source of them. 15 $ for a resistor make me think about. I mean, I have other cheap resistors between my DAC and the speakers (not many around 5). Unless I don't replace them all, I am not sure it would be worth it.
Thanks,
D.
Lucas, I saw that, I was looking for a cheaper source of them. 15 $ for a resistor make me think about. I mean, I have other cheap resistors between my DAC and the speakers (not many around 5). Unless I don't replace them all, I am not sure it would be worth it.
Thanks,
D.
Vishay Z-Foils. TX2575. I dont think you'll regret the investment, especially if there is something else in their first to try.
Direct from texas instruments I get them for for about 10.00 each before shipping and handling. That may be another $10.00 for that. Maybe more for asia shipping. Anyone can order from there I think, even just 1 resistor.
Use caddock for 220k, 1M anything else IMO.
Direct from texas instruments I get them for for about 10.00 each before shipping and handling. That may be another $10.00 for that. Maybe more for asia shipping. Anyone can order from there I think, even just 1 resistor.
Use caddock for 220k, 1M anything else IMO.
I am using PRP Audio grade Metal Film resistors, which cost 30c each. The price/quality ratio for these is unsurpassed, period.
I think it's pretty ridiculous to spend that kind of money on resistors before you've let the circuit settle in - like using $30 silver/gold film caps in a power supply or gold wiring.
For me, such resistors would be a potential tweak for a year or so down the road, if I felt the need. That way you can hear for yourself if you've wasted you're money or not.
I think it's pretty ridiculous to spend that kind of money on resistors before you've let the circuit settle in - like using $30 silver/gold film caps in a power supply or gold wiring.
For me, such resistors would be a potential tweak for a year or so down the road, if I felt the need. That way you can hear for yourself if you've wasted you're money or not.