Increasing the bias current on my NAKSA80s & Sorayas ... seems to be a good thing!!
In my system (actively tri-amped Maggies driven by 2x NAKSA 80s (for mids & ribbons) and a Soraya on the bass panels), I suffer from a lack of soundstage depth. Unfortunately, this is due to the (lack of) distance between the Maggies and the front wall - which is caused by the room dimensions. I may be able to improve soundstage depth by treating the front wall with diffusion products but, in the meantime, with Hugh's advice, I decided to increase the biasing of the amps.
Hugh told me that increasing bias delivers more soundstage depth and suggested I could double the bias currents without causing any problem for the output devices - this means:
* 70mv across the 0.33 ohm res (R22) on the NAKSA 80s (instead of 35mV), and
* 120mV across the emitters of the outermost output transistors on the Soraya modules (instead of 60mV).
As increased bias means increased heat into the heatsinks, I decided to go slightly conservative - so a bit less than twice the stock bias:
* 60mv across the 0.33 ohm res on the NAKSA 80s, and
* 110mV for the Soraya modules.
I'm hoping my heatsinks will be able to cope with this in summer - as the result is delightful! 🙂 Yes, a bit more depth ... but also more 'body' to the sound! 😀
So this is something that AKSA owners might like to try out.
Andy
In my system (actively tri-amped Maggies driven by 2x NAKSA 80s (for mids & ribbons) and a Soraya on the bass panels), I suffer from a lack of soundstage depth. Unfortunately, this is due to the (lack of) distance between the Maggies and the front wall - which is caused by the room dimensions. I may be able to improve soundstage depth by treating the front wall with diffusion products but, in the meantime, with Hugh's advice, I decided to increase the biasing of the amps.
Hugh told me that increasing bias delivers more soundstage depth and suggested I could double the bias currents without causing any problem for the output devices - this means:
* 70mv across the 0.33 ohm res (R22) on the NAKSA 80s (instead of 35mV), and
* 120mV across the emitters of the outermost output transistors on the Soraya modules (instead of 60mV).
As increased bias means increased heat into the heatsinks, I decided to go slightly conservative - so a bit less than twice the stock bias:
* 60mv across the 0.33 ohm res on the NAKSA 80s, and
* 110mV for the Soraya modules.
I'm hoping my heatsinks will be able to cope with this in summer - as the result is delightful! 🙂 Yes, a bit more depth ... but also more 'body' to the sound! 😀
So this is something that AKSA owners might like to try out.
Andy
Thanks Andy!
Felipe, I have taken the Maya to 150mA (54mV across the two 0.18R source resistors, TP1 to TP2) and it does grow a little scale. But I'm reluctant to take it more because with 65V rails, the heat dissipation is starting to move up, particularly in a hot climate.
You could take it up to 200mA for a brief period if you wish, that is, 72mV. But it's dissipating 26W at each of the two outputs into a 0.27C/watt heatsink, and this would mean 15C above ambient at idle, and considerably more heat if it's working hard.
You Maya is going well, Felipe? Are you as happy as you were a month back? I do hope so...... today I visited a Melbourne customer, Russell, who has a 125W Maya on recent Harbeth large, stand speakers - beautiful speakers, absolutely wonderful. We compared the Maya with a DIY 2A3 SET amp with 3.5W per channel, and we come to the same conclusion that the Maya was almost as musical but a little more clear than the SET, with a similar sound stage, very good depth. I was pretty happy about that.....
Cheers,
Hugh
Felipe, I have taken the Maya to 150mA (54mV across the two 0.18R source resistors, TP1 to TP2) and it does grow a little scale. But I'm reluctant to take it more because with 65V rails, the heat dissipation is starting to move up, particularly in a hot climate.
You could take it up to 200mA for a brief period if you wish, that is, 72mV. But it's dissipating 26W at each of the two outputs into a 0.27C/watt heatsink, and this would mean 15C above ambient at idle, and considerably more heat if it's working hard.
You Maya is going well, Felipe? Are you as happy as you were a month back? I do hope so...... today I visited a Melbourne customer, Russell, who has a 125W Maya on recent Harbeth large, stand speakers - beautiful speakers, absolutely wonderful. We compared the Maya with a DIY 2A3 SET amp with 3.5W per channel, and we come to the same conclusion that the Maya was almost as musical but a little more clear than the SET, with a similar sound stage, very good depth. I was pretty happy about that.....
Cheers,
Hugh
Would this also work for the older aksa70 as I have two in my bedroom system that lack a little live and sparkle in general.
No, do not do this with the NAKSA 70. Bias control is not to precise as the 80 and the Maya, and taking it out to high quiescent is dangerous. Best left around 70mA, no more.
Hugh
Hugh
Sjef,
I sold the NAKSA 70 only about a year, and had a couple of failures as a result of enthusiastic fiddling with the bias. I thought about it over quite a bit of time, did some tests, and found that over about 110mA I could lose temperature control of the output stage in some situations, particularly hot weather and strong bass. It prompted me to release a major redesign, the NAKSA 80, and for nearly three years have had no failures of that model. I am extremely cautious about temperature compensation on my amps now, and take pains to ensure that increasing temperatures in arduous conditions the quiescent REDUCES, to prevent thermal runaway. I have honed this approach over a couple of years and my new amps have sophisticated temperature control, using direct sensors on the output device, to measure heat very quickly and react immediately. I have found this issue critical in powerful amps using very large mosfets, like my Maya, which uses very large 480W devices featuring huge transconductance, that is, small increases in gate voltage result in huge increases in drain source. Thermal management in large amplifiers is similar to controlling a nuclear reactor, get it wrong, and there is a lot of smoke and fire!
If you are worried about the top end of the amp, which is actually very good on the N70, examine the other areas of the system, particularly the tweeters on your speaker, or perhaps the source.
Cheers,
Hugh
I sold the NAKSA 70 only about a year, and had a couple of failures as a result of enthusiastic fiddling with the bias. I thought about it over quite a bit of time, did some tests, and found that over about 110mA I could lose temperature control of the output stage in some situations, particularly hot weather and strong bass. It prompted me to release a major redesign, the NAKSA 80, and for nearly three years have had no failures of that model. I am extremely cautious about temperature compensation on my amps now, and take pains to ensure that increasing temperatures in arduous conditions the quiescent REDUCES, to prevent thermal runaway. I have honed this approach over a couple of years and my new amps have sophisticated temperature control, using direct sensors on the output device, to measure heat very quickly and react immediately. I have found this issue critical in powerful amps using very large mosfets, like my Maya, which uses very large 480W devices featuring huge transconductance, that is, small increases in gate voltage result in huge increases in drain source. Thermal management in large amplifiers is similar to controlling a nuclear reactor, get it wrong, and there is a lot of smoke and fire!
If you are worried about the top end of the amp, which is actually very good on the N70, examine the other areas of the system, particularly the tweeters on your speaker, or perhaps the source.
Cheers,
Hugh
That might explain why one of the amps runs warmer then the other. I'm using them in an active threeway system for mids and highs. In fact I like their treble the best as it keeps good control over a pair of old ess sir motion transformers with can get a little harsh on lesser amps. I was more talking about midrange, while pretty good just not as lush as a real classe A. Nevertheless highly enjoyable.
What is your view for the long term regarding temperature management. I tend to keep my equipment for periods of more then 10 sometimes even 25 years. Should I be worried about it?
What is your view for the long term regarding temperature management. I tend to keep my equipment for periods of more then 10 sometimes even 25 years. Should I be worried about it?
Hi Sjef,
Long term should be fine if you ensure all quiescents are set at 70mA.
If one is hot and the other cool, they are unequal, and should be corrected.
OTOH, I have later, better models now; NAKSA 80 has better top end, and SAKSA 80 has better midrange, but of course more designs cost you (and me!) more money.......
Despite my constant attempts, I cannot apologise for my latest amps sounding better than the older models! It's just life...... like buying a new car, and peeling an onion.
Ciao,
Hugh
Long term should be fine if you ensure all quiescents are set at 70mA.
If one is hot and the other cool, they are unequal, and should be corrected.
OTOH, I have later, better models now; NAKSA 80 has better top end, and SAKSA 80 has better midrange, but of course more designs cost you (and me!) more money.......
Despite my constant attempts, I cannot apologise for my latest amps sounding better than the older models! It's just life...... like buying a new car, and peeling an onion.
Ciao,
Hugh
Hi Hugh,
It's not that one amp run hot. In fact it only runs warm, around 30 degrees or fo while the other one remains at almost room temperature. Both amps are virtually doing nothing. They drive a 93db / 16 ohm midrange driver from 150hz and a 95db / 8 ohm tweeter from 3500hz on up.
No need to apologize for anything, offcoarse there is always something better. I don't have any problems with that.
It's not that one amp run hot. In fact it only runs warm, around 30 degrees or fo while the other one remains at almost room temperature. Both amps are virtually doing nothing. They drive a 93db / 16 ohm midrange driver from 150hz and a 95db / 8 ohm tweeter from 3500hz on up.
No need to apologize for anything, offcoarse there is always something better. I don't have any problems with that.
I recently did something similar but took it all the way to its maximum: full Class A. I had an old Apex FX8 (Bimo mod) Class AB that uses lateral FETs that was gathering dust on the shelves. Since I had made a new Class A case and PSU, I decided to install the FX8, but run it in full Class A. Normally bias would be set around 80mA to 120mA. I cranked the bias adjust pot up and up and up. I could hear the trafo start to hum, yikes so backed off to 1amp. It ran fine and sounded better than it ever had. A few days later, I took it to 1.25 amps. Now there is stays and I must say I enjoy the sound. I am told by Bimo that I should be able to get 16w at 8ohms in Class A with 24v rails. This was a successful and safe experiment as these were lateral FETs. I can't say I would do this with my hexFET Class AB amp. It probably would have run away and magic smoke would have resulted. Anyhow, wanted to chime in with my limited experience of pushing an amp to the max bias. Heat = Fun 🙂
Funny thing I bought the Naksa's as a cool running replacement for my claase A mid amps and single ended truode tweeteramps.
Funny thing I bought the Naksa's as a cool running replacement for my claase A mid amps and single ended triode tweeteramps.
Sorry, late Sjef, my daughter was married in Manhattan today and I watched on Skype!!
Yes, the NAKSAs are all Class AB but designed to sound a bit Class A due to the harmonic profile created in the first and output stages.
It is actually a huge tribute to them that you consider them in such rare company; a single ended triode is my idea of the perfect musical sound, but I has many, many technical issues, particularly transformers, filaments and low power.
Hugh
Yes, the NAKSAs are all Class AB but designed to sound a bit Class A due to the harmonic profile created in the first and output stages.
It is actually a huge tribute to them that you consider them in such rare company; a single ended triode is my idea of the perfect musical sound, but I has many, many technical issues, particularly transformers, filaments and low power.
Hugh
To be honoust the single ended triode sound a bit better but they produce a lot of heat, consume a lot power and not to mention expensive tube replacement, specially in a system that's being used by the whole family and therefore runs many hours a day. The Naksa's do share some of it's sound character, at least more then most transistor amps. I stopped using the triode amps after the second 500 euro tube replacement. The Naksa runs cooler and are more worry free and cheaper in the long run.
I have found that the same amp can run hot or cool, solely on the speaker its connected to.
I have a Maya connected to a Sonus Faber Amati ( 80Kg monster speaker 4ohm/92dB) and it probably runs at 10 to 15 above ambient temperature.
On the other hand, i have another Maya connected to a Sonus Faber Elipsa Auditor ( mini monitor speakers 4ohm/89db) and this amp GETS HOT ! I mean very HOT like 30 degrees above ambient !
I have a Maya connected to a Sonus Faber Amati ( 80Kg monster speaker 4ohm/92dB) and it probably runs at 10 to 15 above ambient temperature.
On the other hand, i have another Maya connected to a Sonus Faber Elipsa Auditor ( mini monitor speakers 4ohm/89db) and this amp GETS HOT ! I mean very HOT like 30 degrees above ambient !
Yes, Andy, and it's very interesting.......
It tells me that the crossover on the first speaker is probably series, and the second, shunt, and maybe also the second speaker has lower sensitivity.
Thanks Filipe, useful information for the amp and the speaker!!
Hugh
It tells me that the crossover on the first speaker is probably series, and the second, shunt, and maybe also the second speaker has lower sensitivity.
Thanks Filipe, useful information for the amp and the speaker!!
Hugh
Sure, b_b ... but the interesting thing is ... that for 2 identical power-output Mayas:
* the larger S-F speaker causes the Maya to run warm, whilst
* the smaller S-F speaker causes it to run hot.
Andy