Hi Jim,
I agree with Hugh, those coolers should be very warm/Hot and the copper pad should be Hot pumping 100W into that sucker.
The ceramic insulator pads work great but you also need goop on both sides of the pads.
Are the original cooler spring/bolt removed from the cooler and replaced with standard fasteners for proper clamping pressure?
I agree with Hugh, those coolers should be very warm/Hot and the copper pad should be Hot pumping 100W into that sucker.
The ceramic insulator pads work great but you also need goop on both sides of the pads.
Are the original cooler spring/bolt removed from the cooler and replaced with standard fasteners for proper clamping pressure?
Yes I have goop on both sides. I'm using the original PC cooler springs. I thought it would be nice to use them . Thinking they'd exert decent pressure. Maybe they springs aren't enough for this application. Shall have a look over the weekend and maybe get some music out of it. Thanks Vunce and Hugh for your replies.
Hugh. Alas my weekend isn't the one I'd hoped for so apart from removing the spring loaded fixings from the clamps I haven't got much further. Hopefully this afternoon shall provide me some more free time. One thing I've noticed/overlooked is my aluminium slab is marginal as a heatsink for the CapMx MOSFETs!
Shall report back when I have updates!
Shall report back when I have updates!
I must ask 2 critical question before starting my AN 4R build. In a dual mono configuration using 4 CRC filter boards and 2 dual power supplies for positive and negative +-20vdc per channel, should I use the 10R NTC thermistor and 0,022uF in every pcb or only on the AN pcbs? What is the value of the led resistor at +-20Volts? I think that 2K2 would be OK.
Thanks a lot
Thanks a lot
So I've changed out the spring clamping system for plain screw down. Redid the goop and I now have consistent temperatures, but the sinks are still warm only. The metal tabs of the MOSFETs are about 50c with a k probe. The copper right next to them is 15C cooler. So I seem to be not getting the heat transfer even with ceramic and goo. Might try some 'soft' isolation pads incase there is something out of flat. Or maybe the coolers are so efficient they are just getting the energy out of the copper very well.
Edit.....I unplugged the fans and within maybe 30secs the copper pads were temp equilibrium with the MOSFETs, and fans back on soon cooles down. I just think these coolers with the fans on even very slow are very efficient. I'm no heat engineer but this is all very interesting.
Maybe I can crank the bias a bit more then? 0.1R?!
Edit.....I unplugged the fans and within maybe 30secs the copper pads were temp equilibrium with the MOSFETs, and fans back on soon cooles down. I just think these coolers with the fans on even very slow are very efficient. I'm no heat engineer but this is all very interesting.
Maybe I can crank the bias a bit more then? 0.1R?!
Hi Meta,
For a 3mA LED current with 20V, about 5k6 is fine for the resistor. 2k2 will give you about 8.2mA, a little high.
Jim,
Congratulations! 50C on the metal side of the mosfet is fine.
I'm pretty sure you have 750mV across the two source resistors (check this), so to set idle quiescent the relationship is 0.750/(2x0.1)=3.75A.
If anything, for a 4R AN this is more than enough.
HD
For a 3mA LED current with 20V, about 5k6 is fine for the resistor. 2k2 will give you about 8.2mA, a little high.
Jim,
Congratulations! 50C on the metal side of the mosfet is fine.
I'm pretty sure you have 750mV across the two source resistors (check this), so to set idle quiescent the relationship is 0.750/(2x0.1)=3.75A.
If anything, for a 4R AN this is more than enough.
HD
The leds have been purchased from mouser as per Andy's 4R BOM (TLHR4200 red led, LTL-4231N-1 green led). The have Vf 2V and 2,1V. If 10mA and 20mA. Using the formula (V-Vf)/If the resistors values are (20-2)/0,1=1K8 and (20-2,1)/0,2=895R. The NTC 47R from the 4R BOM must not be used for ground lift as X has clearly stated. He uses the 8R 8D20 series.
Shall I use 10R NTC and 0,022uF in every board of my power supplies in dual mono or it will be ok to add them on the AN 4R boards only? I use 4 CRC filter boards and 2 dual power supplies.
Thanks a lot
Shall I use 10R NTC and 0,022uF in every board of my power supplies in dual mono or it will be ok to add them on the AN 4R boards only? I use 4 CRC filter boards and 2 dual power supplies.
Thanks a lot
Jim,
For a 4R load, I'd increase the current to 2.2A - a lot more heat of course, but necessary.
380mV across each source resistor is 760mV across the two, close to my estimate.
Accordingly, I'd use 0.380/0.18=2.2A, close enough.
The calculation for peak from 25V rails with 4R resistive about 5A is peak current, so half this for each source resistor and you have 2.5A.
(Twice the quiescent gives you the peak current).
Since with 28.3V rails (sagging a little) your dissipation at EACH output will be 28.3V x 2.5A = 70.5W. This is very high, and your thermals will be problematic, so maybe unless you can reduce the voltage to say 22V max I would go with the 0.18R source resistors for 2.11A, ie 59.7W. Still high, but doable.
HD
For a 4R load, I'd increase the current to 2.2A - a lot more heat of course, but necessary.
380mV across each source resistor is 760mV across the two, close to my estimate.
Accordingly, I'd use 0.380/0.18=2.2A, close enough.
The calculation for peak from 25V rails with 4R resistive about 5A is peak current, so half this for each source resistor and you have 2.5A.
(Twice the quiescent gives you the peak current).
Since with 28.3V rails (sagging a little) your dissipation at EACH output will be 28.3V x 2.5A = 70.5W. This is very high, and your thermals will be problematic, so maybe unless you can reduce the voltage to say 22V max I would go with the 0.18R source resistors for 2.11A, ie 59.7W. Still high, but doable.
HD
Hugh, apologies. I didn't see this post and I was relying on the forum alert. Remiss of me.
Thanks for that, very helpful. I'm still learning the maths, and when I need to multiplying by 2 for the output devices etc is still sinking in.
I actually got the amp in the system playing music for the first time last night. Don't ask me where the last two weeks went?! Best intentions and all that.
Sounds very nice on first acquaintance. Only got maybe half an hour of music but then it played TV with the wife for a couple hours. No issues. Offset about 5 to 7 MV each channel. No turn on/off thumps.
Gain is higher than I expected so that's nice. Not too high though so I can still wind the Salas DCG3 preamp (with 3x gain) up a decent bit. I need to review my CapMx PSU. It runs pretty hot on the slab of aluminium as heatsink. I used a couple 80mm fans aimed at it and helps a lot but I can hear them. Can't hear the main 120mm fans on the CPU coolers. On that note I think the added dissipation of going to 0.18R source resistors should be no problem for this active cooling set up.
Thanks for that, very helpful. I'm still learning the maths, and when I need to multiplying by 2 for the output devices etc is still sinking in.
I actually got the amp in the system playing music for the first time last night. Don't ask me where the last two weeks went?! Best intentions and all that.
Sounds very nice on first acquaintance. Only got maybe half an hour of music but then it played TV with the wife for a couple hours. No issues. Offset about 5 to 7 MV each channel. No turn on/off thumps.
Gain is higher than I expected so that's nice. Not too high though so I can still wind the Salas DCG3 preamp (with 3x gain) up a decent bit. I need to review my CapMx PSU. It runs pretty hot on the slab of aluminium as heatsink. I used a couple 80mm fans aimed at it and helps a lot but I can hear them. Can't hear the main 120mm fans on the CPU coolers. On that note I think the added dissipation of going to 0.18R source resistors should be no problem for this active cooling set up.
Attachments
Hi Hugh,
As you know, I built the 4R version - as the Maggie drivers I was using them on were 2 and 3.1 ohms.
However, the 3a quiescent current (resulting from +/-22v DC rails and the specified 0.12R Source resistors) puts a lot of heat into the heatsinks.
So I'm wondering what the effect would be of increasing these resistors slightly?
Is it true that ... the higher the quiescent current is ... the better able the amp is at driving a 4 ohm load? (And 3a is happy driving 2 ohms!)
The output measurements I took (with R15 & R16 on the AN 4R schematic = the specified 0.12R) were as follows
-----------------------------------------------------------
6.2 ohm load:
Sine wave clips at 40v pp (top peak!)
Red LED goes on at 38v pp ... so it indicates that clipping is about to happen!
Power is 31w, at clipping
Current is 3.22a
3.7 ohm load:
Sine wave clips at 37.5v pp (bottom peak!)
Red LED goes on at 36v pp ... so it indicates that clipping is about to happen!
Power is 47w, at clipping
Current is 5.06a
3.1 ohm load:
Sine wave clips at 36v pp (bottom peak)
Red LED goes on at 36v pp ... so there is no warning!
Power is 52w, at clipping
Current is 5.8a
2.5 ohm load:
Sine wave clips at 29v pp (bottom peak)
Red LED goes on after clipping starts - so no warning!
Power is 42w, at clipping
Current is 5.8a
2.1 ohm load:
Sine wave clips at 24v pp (bottom peak)
Red LED goes on after clipping starts - so no warning!
Power is 34w, at clipping
Current is 5.7a
-----------------------------------------------------------
I proved the AN 4R circuit was stable into a 2 ohm load (my Maggie ribbons) - but - now that my (new) spkrs offer a 4 ohm load (not 2 ohm/3.1ohm) - could I use, say, 0.15R resistors for R15 & R16, in place of the existing 0.12R resistors?
This would reduce the quiescent current and therefore the heat load into the heatsinks.
I presume it would make the amp less able to cope with a 2 ohm load ... but are there any other negative results of increasing the Source resistors? (As driving a 2 ohm load is no longer important.)
Thanks.
Andy
I guess there is more to just the Iq for the 4R version as it isn't just a case of swapping the source resistors. There are quite a few changes to smaller resistors and electrolytic caps in the 4r BOM v 8r BOM. Would be interesting to know what these changes to the components actually does to the behaviour of the circuit. Just to further my understanding.
Yes Andy, 0R18 rather than 0R15 would be just fine.
Actually, I'd probably take it to 0R20 resistor if you can get 'em.
Reduces the output device heat a little more and your easier load will cope well.
I calculate about your quiescent would be 1.9A for 0R20 resistors, and your max current would be 3.8A, which with 4R load would be a peak voltage of 15.2Vpk, which corresponds to 28.9W of audio.
Hugh
Actually, I'd probably take it to 0R20 resistor if you can get 'em.
Reduces the output device heat a little more and your easier load will cope well.
I calculate about your quiescent would be 1.9A for 0R20 resistors, and your max current would be 3.8A, which with 4R load would be a peak voltage of 15.2Vpk, which corresponds to 28.9W of audio.
Hugh
The clipping indicator was added as a simple feature for use with the 8R design. That it works well at lower impedances does not surprise me.
To reveal any clipping, you should connect a CRO during your listening tests, since the clipping is always on the negative half cycle. The current saturation at clip is limited to twice the quiescent current, which is unlikely to damage your speakers.
HD
To reveal any clipping, you should connect a CRO during your listening tests, since the clipping is always on the negative half cycle. The current saturation at clip is limited to twice the quiescent current, which is unlikely to damage your speakers.
HD