I always go for the highest rated transformer I can afford (maybe up to 500VA) just in case at a later stage you want to build something else or tweak the current amp you have to higher bias current.
Otherwise you might end up having to buy another transformer in the future if you plan to do any of the above.
Well that is my opinion on it.
Otherwise you might end up having to buy another transformer in the future if you plan to do any of the above.
Well that is my opinion on it.
I always consider the followings:
- Transformer suppliers are not always ideal.
- Transformers show different levels of unwanted power losses.
- Considering A and B above, assume the power loss of about 20%.
- Follow up designers' rule of thumb: transformer's VA to be two times of circuit draw.
[/list=A]
So, my case with F5 of +/- 24V and 1.3A is:
Per channel, (2/0.8)*(24+24)*1.3= 156VA
For both channel, total minimum 300VA
🙂
Isn't the F5 able to drive loads lower than 8 ohms and so, to reach twice its power?
May be this could explain why Nelson oversized the transformer.
May be this could explain why Nelson oversized the transformer.
I would assess the transformer requirement on three different criteria.
1.) the ability to supply ClassA bias current continuously without heating the the manufacturers rated maximum temperature.
2.) the general requirement for a ClassA transformer to be rated to somewhere between 6times and 10times the maximum ClassA output power.
3.) the ClassAB requirement for the transformer to be rated somewhere between 1times and 2times the maximum ClassAB output power.
for:
1.) 1.3A bias current requires the transformer to be rated @>=2.6A continuously to match the manufacturers maximum temperature rating.
Allow double this to run the transformer cooler continuously, i.e. ~5A, about 240VA
2.) for 25W maximum ClassA output the range of rating could be between 150VA and 250VA
3.a) for 50W into 4ohm, a transformer rated between 50VA and 100VA would do.
3.b) for 100W into 2ohm, a transformer rated between 100VA and 200VA would do.
Note that the ClassAB requirements are the least onerous.
The maxima for cases 1. & 2. are very similar.
For two channels operation it appears that 480VA or 500VA is about right.
1.) the ability to supply ClassA bias current continuously without heating the the manufacturers rated maximum temperature.
2.) the general requirement for a ClassA transformer to be rated to somewhere between 6times and 10times the maximum ClassA output power.
3.) the ClassAB requirement for the transformer to be rated somewhere between 1times and 2times the maximum ClassAB output power.
for:
1.) 1.3A bias current requires the transformer to be rated @>=2.6A continuously to match the manufacturers maximum temperature rating.
Allow double this to run the transformer cooler continuously, i.e. ~5A, about 240VA
2.) for 25W maximum ClassA output the range of rating could be between 150VA and 250VA
3.a) for 50W into 4ohm, a transformer rated between 50VA and 100VA would do.
3.b) for 100W into 2ohm, a transformer rated between 100VA and 200VA would do.
Note that the ClassAB requirements are the least onerous.
The maxima for cases 1. & 2. are very similar.
For two channels operation it appears that 480VA or 500VA is about right.
Misreading exepted, Nelson says 500VA for a 100w class AB
http://www.passdiy.com/pdf/articles/powersupply.pdf (2001)
I do own a Yamaha professional series the Xformer of which is rated 750va for 2x80w.
http://www.passdiy.com/pdf/articles/powersupply.pdf (2001)
I do own a Yamaha professional series the Xformer of which is rated 750va for 2x80w.
Had a question all typed up but edited in favor of a different question.
So if we have a bias current of X how does this effect the ClassA heat dissipation of the mosfets? Whats the formula?
Also, for this specific amount of VA and assuming zero output power, just running idle, how would you rate a transformer? Do you over rate for running idle in ClassA as well or do you just rate the transformer on a straight 1:1 relationship for the idle heat dissipation?
I guess I am just trying to isolate one part of the amp power/transformer power relationship. We always talk about output power and how it goes into deciding transformer size but I want to know how that automatic given of idle heat dissipation effects our transformer needs.
Uriah
So if we have a bias current of X how does this effect the ClassA heat dissipation of the mosfets? Whats the formula?
Also, for this specific amount of VA and assuming zero output power, just running idle, how would you rate a transformer? Do you over rate for running idle in ClassA as well or do you just rate the transformer on a straight 1:1 relationship for the idle heat dissipation?
I guess I am just trying to isolate one part of the amp power/transformer power relationship. We always talk about output power and how it goes into deciding transformer size but I want to know how that automatic given of idle heat dissipation effects our transformer needs.
Uriah
I thought Nelson provided a great explanation of how to size transformers
for class A amplifiers in his Zen 5 paper...
for class A amplifiers in his Zen 5 paper...
sizing transformers
My favoriite is still
"calculations we don't need no stinking calculations we just make em really big"
My favoriite is still
"calculations we don't need no stinking calculations we just make em really big"
bobodioulasso said:
I have approximated the minimum 300VA for both channels, considering pure class A operation with +/-24V and 1.3A bias in post#3962, i.e. 25W@8-ohm load, 12.5W@4-ohm load and 6.3W@2-ohm load . . .
Above these class A output power, the amp will operate as class B, e.g. up to 50W@4ohm-load and 100W@2ohm-load . . .
And, if I want to get the maximum output power up to 100W@2-ohm load, I need pk-pk ac current of 20A, and I have to reconsider the power transformer VA rating, by adding the dc current drain (average value of the half wave signal) of the class B push-pull to the 1.3A class A bias. According to my approximation in this case, total dc current has to be about 1.3+0.7=2A. Then, using the same way in post#3962,
Per channel: (2/0.8)*(24+24)*2= 240VA
For both channels: total minimum 480VA
Actually, I'm using 500VA transformer for my F5 . . .
🙂
Okay, I saw a guy posting about modding his amp (forget which one, some super commercial brand) by removing some resistors and replacing them with precision current sources. He used LM334 TO-92 packages with current setting resistors in place of a few resistors. Claimed it was great for improving the cymbals sound.
What do you guys think? Hogwash or room for experimentation?
I dont know if this was used in the signal or not, wish I had bookmarked the page.
Uriah
Okay, I found the reference
http://www.drmaudioht.com/BASIC_M1_CurrentSource.htm
Not as I remembered it but its interesting anyway.
What do you guys think? Hogwash or room for experimentation?
I dont know if this was used in the signal or not, wish I had bookmarked the page.
Uriah
Okay, I found the reference
http://www.drmaudioht.com/BASIC_M1_CurrentSource.htm
Not as I remembered it but its interesting anyway.
F5 + quad
Have anyone of you nice people tried the F5 with the ESL-63 or the original Quad electrostatic?
bobodioulasso said:
Have anyone of you nice people tried the F5 with the ESL-63 or the original Quad electrostatic?
In the manual Nelson says:
"In spite of the thermal compensation in the circuit, you should assume that there will be drift as the heat sink temperature
rises, and you will need to readjust the values over the course of an hour or two. Usually it is best to start out bias
adjustment low, at maybe 0.4 mV across R11 and R12 until the amp is warmed up a bit"
Is the drift because P1 and P2 drift in resistance as they warm up from the heat of the heatsink? If not, why?
Uriah
"In spite of the thermal compensation in the circuit, you should assume that there will be drift as the heat sink temperature
rises, and you will need to readjust the values over the course of an hour or two. Usually it is best to start out bias
adjustment low, at maybe 0.4 mV across R11 and R12 until the amp is warmed up a bit"
Is the drift because P1 and P2 drift in resistance as they warm up from the heat of the heatsink? If not, why?
Uriah
You start around 0.4 V, not mV.
The drift is due to temperature change in the semiconductors.
😎
The drift is due to temperature change in the semiconductors.
😎
Thanks Mr. Pass. And thanks for the clarification on .4V vs .4mV, thats a pretty big difference.
Uriah
Uriah
I usually double this rating for a 2x100w class AB amp (bottom in multi-amplification). Though, who will use F5 on a 2 ohm load? Is'nt the limiting current circuit set too low for this ( i do not remember, as i do not use any)?
1kVA for 100W+100W ClassAB is unusually high. I don't think there are many here who would recommend that VA to W ratio and few builders who would follow that advice.bobodioulasso said:
It is only sensible to do if you have 100W/8Ohms amps that doubles and doubles each time you divide the load by 2, i.e 200W/4Ohms, 400W/2Ohms and so on.
AndrewT said:
I do not give any advice, just telling what i do.
For a 2x 100w/8 ohms capable of 2x400w/2 ohms i do not think i oversize so much the transfo rating. ( See commercial P.A. products)
In the case of the F5/2ohms capable, of coarse , it would be oversized.
As a rule of thumb, when Nelson recommends 600Va, i prefer to follow his instructions rather than to loose myself into uncertain calculations.
Though, Nelson's well known generosity, certainly allow us some margin in our transformer rating choice ( depending on availability and price).