Anand, Scott,
I have spent some time on the AN39 using a lateral n type for the upper device.
Results are mixed.
IF we use a double die lateral (the 16A 200V 250W EXW20N20 Exicon device) AND a FQA36P150 (36A 150V 280W Fairchild as the pmos variable CCS, we can raise the second harmonic about 5 dB to about -74dB (at 12.5W into 8R, +20dBU). But the third harmonic will also rise, to about -79dB, a little higher.
The difference in the transconductance here is considerable; gm from 1.4S (lateral) to 19.5S (hexfet) and this range is about the best you can achieve. With any other combination than a nlateral and a pmos hexfet we reduce gm differences, and therefore the creation of asymmetric harmonics (that is, even order harmonics) will be less, so it's the best we can get at present. BTW, reducing the gm differences will reduce the THD overall too.
I have spent many hours on Spice and I can't see a way around. By decreasing the global feedback by increasing nested feedback does not seem to much change the profile. The fact is the topology is so linear that creating higher levels of low order harmonics, in particular H2, is tricky........
If you like a bit of highish H2 and lower H3 this amplifier may not satisfy. Scott, you are in this category, and I would suggest some of the Pass amps with higher damping factor might be more suitable. I'm sorry the AN does not suit your needs; but this is to be expected, not even people like their very low distortion sound, a lot of people like a strong mix of harmonics, leading them towards tube amplifiers.
Cheers,
Hugh
I have spent some time on the AN39 using a lateral n type for the upper device.
Results are mixed.
IF we use a double die lateral (the 16A 200V 250W EXW20N20 Exicon device) AND a FQA36P150 (36A 150V 280W Fairchild as the pmos variable CCS, we can raise the second harmonic about 5 dB to about -74dB (at 12.5W into 8R, +20dBU). But the third harmonic will also rise, to about -79dB, a little higher.
The difference in the transconductance here is considerable; gm from 1.4S (lateral) to 19.5S (hexfet) and this range is about the best you can achieve. With any other combination than a nlateral and a pmos hexfet we reduce gm differences, and therefore the creation of asymmetric harmonics (that is, even order harmonics) will be less, so it's the best we can get at present. BTW, reducing the gm differences will reduce the THD overall too.
I have spent many hours on Spice and I can't see a way around. By decreasing the global feedback by increasing nested feedback does not seem to much change the profile. The fact is the topology is so linear that creating higher levels of low order harmonics, in particular H2, is tricky........
If you like a bit of highish H2 and lower H3 this amplifier may not satisfy. Scott, you are in this category, and I would suggest some of the Pass amps with higher damping factor might be more suitable. I'm sorry the AN does not suit your needs; but this is to be expected, not even people like their very low distortion sound, a lot of people like a strong mix of harmonics, leading them towards tube amplifiers.
Cheers,
Hugh
Hello my friends I am gathering parts for the 4R AN39 version. What I would like to ask is how critical and audible it is to put carbon resistors in these resistors 820Ω-22K-100gate resistors -10Ω zobel -Rg ..... when the existing ones have a very high value in ppm about 450 (5%)
Won't the temperature affect their value in ohms enough?
Mouser parts
588-OM8215E-R58
279-CFR100J22K 22K carbon
71-CPF110R000FKEE6 10R carbon
660-SPR1CT52R101J 100R carbon
Anyone else with better parts Mouser codes?
Attachments
Last edited:
Hugh,
Thanks for diving in a playing with the sims,
Guys do not get me wrong here the Nirvana is a fantastic sounding amp
Extremely linear and clean sounding
Dead quiet,
Hugh is correct there is no doubt that my brain is accustom to more H2 harmonics
To the point that I notice immediately when It is not there, probably has something to do with fact that I have run Tube amps for years.
I would encourage you guys that are on the fence about building Nirvana to give it a go well worth the trouble.
Thanks to Hugh,and X for all the hard work
In bringing this amp to us.
We are truly blessed to have such a fantastic community to share in our passion!!
Scott
Thanks for diving in a playing with the sims,
Guys do not get me wrong here the Nirvana is a fantastic sounding amp
Extremely linear and clean sounding
Dead quiet,
Hugh is correct there is no doubt that my brain is accustom to more H2 harmonics
To the point that I notice immediately when It is not there, probably has something to do with fact that I have run Tube amps for years.
I would encourage you guys that are on the fence about building Nirvana to give it a go well worth the trouble.
Thanks to Hugh,and X for all the hard work
In bringing this amp to us.
We are truly blessed to have such a fantastic community to share in our passion!!
Scott
The carbon resistors have the biggest impact on the sound through how they impact the harmonic distortion profile when used at the feedback resistor position. I showed this in the Aksa Lender thread and the front end of the AN is based on the Aksa Lender preamp. Carbon film will give a slightly higher overall THD than metal thin film but with higher second harmonic vs third harmonic. Whereas metal thin film gives lowest absolute THD but may have higher third harmonic, possibly equivalent in magnitude as second harmonic. This mostly has to do with the the relatively high signal levels at the feedback resistor and its role as feedback. In any case, do not use thick film (often larger flameproof power resistors are this variety) or wire wound resistors here. The thick film will increase overall THD significantly (10x more). Source or emitter resistors with thick film in the output stage are also bad.
The gate snubber may have an impact as well but I have not investigated the difference. It can’t hurt to use carbon there. Carbon composition “pulse” resistors, I have heard are “sweet” sounding when used as snubbers. This is purely anecdotal from a friend who specializes in all FET amps.
Regarding matching the BJTs - with Hfe numbers in the +/-3 in the last digit should be fine. If you have a perfect match even better. Also look at voltages between base and collector.
Last edited:
Interesting project...I ordered some PCB's and I will build the 4R version...
Did anyone try in the position of the upper mosfet (n-mos) any of the following?
IRFP140R
IRFP150N
FQL40N50
I didn't go thru the whole topic yet, it's huge...I know they are maybe different than those in discussion here, but these types I'm pretty sure I can get NOS, of certified quality...they were made by Harris and International Rectifier.
If I get some answer I'm not asking just about this amplifier, but the use of these mosfets in class A amps; just to know if it's worth trying them. I haven't yet an option for the P-channel...just waiting to see what I can use from the above - if any - and I will search them after...
Did anyone try in the position of the upper mosfet (n-mos) any of the following?
IRFP140R
IRFP150N
FQL40N50
I didn't go thru the whole topic yet, it's huge...I know they are maybe different than those in discussion here, but these types I'm pretty sure I can get NOS, of certified quality...they were made by Harris and International Rectifier.
If I get some answer I'm not asking just about this amplifier, but the use of these mosfets in class A amps; just to know if it's worth trying them. I haven't yet an option for the P-channel...just waiting to see what I can use from the above - if any - and I will search them after...
Thank you! In this case, at least IRFP140/9140 gets out of the list, the 9140 has a transconductance of 5...I guess also the N-channel 140 will have also smaller transconductance. But FQL40N50 shows in datasheet from On-Semi a typical transconductance of 29 Siemens, given at 20A...but I don't find a matching P-channel model for these. I know there's no complementary devices need here, but still I think they should be in the same league...
EDIT: also IRFP150N has 15S, better but still lower than 29.
That "gate capacitance" is the input capacitance from the datasheets?
EDIT: also IRFP150N has 15S, better but still lower than 29.
That "gate capacitance" is the input capacitance from the datasheets?
Last edited:
Mr Hugh, thank you very much for your clarifications. It is correct to keep the FQA40N25 mosfet in the N channel and to complete P with the IXTQ36P15P? I say this since only the FQA40N25 mosfet has a large number in S.
Also, my power supply consists of 2X18AC and classic rectification with diodes MUR 460 CRC 22,000 uf and CapMx which CapMx has a voltage drop of about 2.3V . According to my calculations I will have a voltage at the output of about 19V at 2.2A. I have the same power supply in the USSA5 amplifier that plays with 1.6A and gives me an output of about 21.3V
Hi, I would like your opinion regarding my calculations and what resistors 0.20Ω or 0.18Ω to put in 19V voltage for 2.2A for version 4R I always talk....
Thanks.
Last edited:
Spiros it is my pleasure...,
Yes FQA40N25 is an excellent for FQZ36P15.
19V at the rails is about right but I think your maximum peak will be just a little over 13V peak into 4R load. Your current should be 2.2A as you say and both source resistors should be 0.18R each.
Your total output should be 21.8W into 4R.
Hugh
Yes FQA40N25 is an excellent for FQZ36P15.
19V at the rails is about right but I think your maximum peak will be just a little over 13V peak into 4R load. Your current should be 2.2A as you say and both source resistors should be 0.18R each.
Your total output should be 21.8W into 4R.
Hugh