Actually I use Project P09 from Rod Elliott, opamp based.
Maybe later I will look for a FET based crossover.
Passive crossovers?
I avoid them for every price in the world!
It is technically not logically to do this task in the power domain, I really prefer to do it in the small signal domain.
And the speakers are a very complex load for crossovers. Only magicians can really make "good" passive crossovers.
Specially in conjunction with the F5 it is imho essential to give the low output impedance of the amp the full control over the speakers eg. attach the speakers directly to the F5.
Exception: one cap in series with a tweeter.
Franz
Last edited:
Ouch, yes your are right, Ramallo: they mixed up source and drain. But as far as I know not important for JFETs.
While checking out the amp, I realized I have overseen, that I cut too much of the already quite low voltage for the input stage (R3/R4 drop) when using too high bias settings.
So, I reduced the idle current down to 0.7A each channel.
Yesterday, I wrote:
Hmmm, I was wrong!
I measure an inbalance at R3 and R4: all boards are showing now a voltage drop of 2.3V on R3 and 4.6V on R4 :-(
What will be the most possible cause?
Poorly selected 2SK170/2SJ74?
Differences in the IRFP9240/240?
Or the mentioned pcb layout bug?
Franz
Hi Franz,
have the same ESP P09 project so it's not a surprise to share your opinion completely. The difference to pasive is really unbelivible (stage, details)!
BTW, for the moment 2 ESP amps (Daeth of Zen, P101) and 1 Hypex (ucd 400) are pushing my speakers. Otherwise, PCBs for F3, F5 and Aleph J-x amp, also some parts, are waiting to have enough time...
Sorry to all for OT!
Hope you solve the problems with your F5 mini very soon!
O.K. I think I got it with my Mini F5:
I set the amps on idle current of 0.63A.
So, I have +-11.55 on the rails.
The heathsinks are now about 20degrees over ambient temperature during operation.
And the sound is as fine as from a full F5.
The only question remaining:
On R3 / R4 the voltage drop is 4.5VDC each.
Why is it higher than in the full F5? Has it anything to do with the fact that I did not build in the current limiting circuit?
I measure at R1/R2 45mV, so the input stage is consuming 4.5mA (Idss of the selected FET's 6.8mA at 10V).
Franz
I set the amps on idle current of 0.63A.
So, I have +-11.55 on the rails.
The heathsinks are now about 20degrees over ambient temperature during operation.
And the sound is as fine as from a full F5.
The only question remaining:
On R3 / R4 the voltage drop is 4.5VDC each.
Why is it higher than in the full F5? Has it anything to do with the fact that I did not build in the current limiting circuit?
I measure at R1/R2 45mV, so the input stage is consuming 4.5mA (Idss of the selected FET's 6.8mA at 10V).
Franz
Last edited:
That voltage depends on transfer characterisitic of the HEXFETs you have (not all HEXFETs are created equal) and voltage drop accross R11 or R12 (which depends on their value and bias current you have chosen).
O.K.
I understand. They need a certain voltage on the gate to get conducting eg to reach the desired operation point.
Yes, the current about R11 and R12, this I understood before
I have now 0.3VDC across 0.47 Ohm = 0.63A
Franz
Have you 0.63A on both (240 and 9240)?, whats is your osffet voltage?.
I can't get the same value for the 9240 and the 240 for a reasonable osffet, I have the 9240 at 0,59VDC and 0,57VDC for the 240 (osffet 20mV). I'm thinking in change the value of the 0.47 Ohm resistor of the 9240 (by added another resistor in parallel of 10 Ohm or more).
Cheers
Toshiba's 2sk1529/2sj200.
They have much lower Vgs at Id=0.63A and they will certainly perform better with such a low Vds (your PS voltage is very low and standard HEXFETs' parasitic capacitacies are playing crazy at Vds=11.2V).
OTOH Toshiba MOSFETs have a bit lower Yfs which means a bit higher THD (that may not necesarilly be bad).
Anyhow, with such a low PS voltage I'd use k1529/j200.
If you use 2SK1530/2SJ201, they have about the same Yfs as the IRFPs.
This is well proven, see my posts near the beginning.
I have used them at +/-12V rails without problems.
My balanced version uses +/-16V.
Again all published in this thread ages ago.
Only advantage of 2SK1529/2SJ201 is slightly lower capacitance, IMHO.
Patrick
This is well proven, see my posts near the beginning.
I have used them at +/-12V rails without problems.
My balanced version uses +/-16V.
Again all published in this thread ages ago.
Only advantage of 2SK1529/2SJ201 is slightly lower capacitance, IMHO.
Patrick
Hello Franz,
The resistors are ok (I measured resistors before mount it), I have low bias on the 240's on both channels, maybe a not perfect matched 9240 vs 240.
I'm using 3x DVM (One at output for osffet)
Thanks
I try to resume the experience with the Mini F5 up to now:
After many different bias settings, I did not find a completely satisfying solution (the measuring equipment is not available at the moment).
The very first run with the Veravox 3S fullrange speakers, driven by the low voltage from a iPod, was the best sounding configuration.
Low volume level, due to the low effiecency, the low gain and the low input voltage. Idle current for this test was 1A.
+-11.5VDC seems defintely too low rail voltage to drive the IRFP9240/240 in a linear matter.
As I have some higher voltage trannies in my stock the next tries will probably be with higher voltage prior considering using different MOSFETs.
Ramallo:
Do you really think this is a major problem? It is within a few percent and could be caused by overall circuits tolerances.
Franz
After many different bias settings, I did not find a completely satisfying solution (the measuring equipment is not available at the moment).
The very first run with the Veravox 3S fullrange speakers, driven by the low voltage from a iPod, was the best sounding configuration.
Low volume level, due to the low effiecency, the low gain and the low input voltage. Idle current for this test was 1A.
+-11.5VDC seems defintely too low rail voltage to drive the IRFP9240/240 in a linear matter.
As I have some higher voltage trannies in my stock the next tries will probably be with higher voltage prior considering using different MOSFETs.
Ramallo:
Do you really think this is a major problem? It is within a few percent and could be caused by overall circuits tolerances.
Franz
there are only 4 active devices in the F5.
All are FETs.
Vgs of each FET can be anywhere in the voltage range specified by the manufacturer.
It's that enormous range of Vgs that results in seeing big differences in bias currents and even bigger differences in output offset.
Matching of half the devices will not improve the matching of the other half of the devices.
All are FETs.
Vgs of each FET can be anywhere in the voltage range specified by the manufacturer.
It's that enormous range of Vgs that results in seeing big differences in bias currents and even bigger differences in output offset.
Matching of half the devices will not improve the matching of the other half of the devices.
Todays configuration / settings of my four channel Mini F5:
- +/- 15.2VDC rail voltage
- 0.8 A bias (0.38V at R11/R12)
This trannies are very small but I have no other choice at the moment.
So I gave it a try and it works.
The trannies are 2x12V 80VA. The calculated DC load on the rectifier is now about 50VA, so I should not overload the trannies imho.
They get about 45 C after one hour.
The sound?
Some kind of better than before.
Give me time for my conclusion.
Franz
- +/- 15.2VDC rail voltage
- 0.8 A bias (0.38V at R11/R12)
An externally hosted image should be here but it was not working when we last tested it.
This trannies are very small but I have no other choice at the moment.
So I gave it a try and it works.
The trannies are 2x12V 80VA. The calculated DC load on the rectifier is now about 50VA, so I should not overload the trannies imho.
They get about 45 C after one hour.
The sound?
Some kind of better than before.
Give me time for my conclusion.
Franz
Last edited:
One transformer to drive two channels.
Hard on the limit, I know.
But I remember a story from Willi Studer: when he visited the laboratory, he touched the transformers.
Could he touch them more than five seconds, he judged the tranny to be overdimensioned.
I know, for Audio purposes, specially for power amplifiers, it is worth to overdimension trannies.
So, my trannies are experimental, I think. Maybe I switch to 15VAC trannies later (more VA...).
Soundwise, it seems that I am on the right track to increase the voltage for the Mini F5 version.
Before I can judge I have to complete the "full" F5 soon to drive the woofers (I will use a 500VA for the full version...).
My woofers are playing up to 860Hz, so it is not easy to judge right know the Mini F5, playing above 860Hz.
Franz
Last edited: