F5 power amplifier

Thanks root2.

May consider your products in the future.

I've been researching this ssr relay and its specifications concerning high inductive transformer startup and zero crossing and all. It talks about resolving the issue with a suppressor diode. (don't know what product to buy nor where to place in circuit) There are also better SSR products more suited for tfo startup. The one I am using is probably not suited for such purpose but it does work well for 1 F5 Power supply. I am not familiar enough with the subject matter so I am tending to go the coil relay route.

what I've tested with SSR relays (120v 40 amp version):
1 relay for 1 f5 ps --->> OK results
1 relay for 2 f5 ps ----> noise problem probably caused by what root2 is stating
2 relays in series for 2 f5 ps ---> noise problem present

I haven't tested 1 relay for each f5 PS but I suspect that this would be OK but I'm doubling the already expensive crydom components.

Also, I have seen higher rated crydom products (H1 series fo which I believe that probably would work).

I will use the ssr for my dac project so its not a lost cause.

thanks all for feedback.
 
crydom ssr hum on tfo resolved

I noticed that the coil relay functionned with no hum and the ssr experienced a hum on the transformers in a dual mono config.

The Live wire was going through the relay and to the 2 tfos in parallel. I changed that to serial and hum was no more when going through the ssr
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
 
sorry fro the bad info

that did not work.

no hum but only half the voltage.

So I'm going back to coils. will try to find an appropraite zero cross ssr relay.

You could still use a SSR so you get Zero crossing and the in parallel with it
run relay, the relay will take longer than 10-20ms to operate but will not be killed by the inrush, now using a Snubber and still having the SSR on switch the relay OFF then the SSR lets say 60mS later should solve cross over hum and inrush.

A good SSR will not be open circuit until one complete cycles and at about 1.4V below zero volts.

Have fun :)
 
Last edited:
psu concerns

Hello,
I need your opinion about a psu.

I'm trying to make a psu for F5 using many components from my box.
the transformer is a 45vac output that give around 64v avec diode bridge.
Using virtual ground each side is near 32v. the bleeders resistor is ok.
the main problematic is on the PI resistor.
As the target is 24v,I have to drop near 8volts if I'm right.
As described in nelson pass f5 manual, 24volt for 6amp continuous and more than 10amp for peak.
In this case, for drop 8volt in 6A, near 1.33 ohms is needed. and using 1.33 ohms in 10amp peak, the voltage drop will be near 13.3 volts... seems too much!!!

So regarding this problem, do you think that I have to use regulator, or using the right resistor value will be ok ?

About regulator, i'm thinking of using lt1083, or see in zen variations #5 as nelson pass recommends.
About resistor, what do you think about the use of bigger capacitors to reduce the problem?

thank you for your help
 
I heard some of us have had some in hand! Maybe at BAF last fall???
At any rate, they are still available ju$$t a few dollars more :D

Stock is running low, a few definitely not enough, some still not enough, a bunch would be adequate, :D

According to Linear Systems, they were at BAF last fall with samples. They told me soon and I told them they've been saying that for at least 5 years. So I don't know what the hold up is.

Samples!?, sound promising
 
You can use 2SJ74GR -- these are around. Best if you buy a bunch and get a matched pair to parallel 2. 2N5460 can be used -- noisier by several dB -- different pinout. 2N5460 Idss ranges from less than 1mA up to 4.5mA, transconductance from just under 1mS to over 5mS. The lower transconductance means that you won't get the very low THD attainable with the original F5 and BL devices, but it's in the same ballpark. DIY is about experimenting.
 
Hi,
Thank Jack for the info and suggestion,

I've also been suggested to use 2SK366 and 2SJ107 as the substitute pair, they have closely match parameters with 2SK170/2SJ74, with higher gm and lower PD.
But sadly this pair also in unobtainium realm.

Well for one stop build, it is ok to have few pair with premium price, but IMHO DIYers need more for trial and error experiment.

As stated earlier by Labjr, Linear System were at BAF last fall with samples. At least now they have something, instead of words for nearly five years, let's hope for the best.

For now I will build a stock F5, next will try other F5's circuits and outputs variaton.

Thanks
 
Higher voltage lower bias

Since I plan to eventually build both a F5 and a gainclone, but having not that much of efective speakers, but with pretty high load ( the speakers are these ), I was thinking about building one nice cased power supply with 2x24V secondaries and use it both for a LM3886 based amp and for a F5 that would see much higher rail voltage compared to the stock 24V, but lower the bias below 1A. What effect would this have on the behavior of the F5 ? From the turbo paper I understood that adapting some resistor values I could do without "turbo". Only question is : what do I get ? I would get an amplifier that goes faster into Class A/B that the stock F5 ? Has anybody tried this without doubling the output MOSFET's ?

Regards, Stefan
 
Moderator
Joined 2002
Paid Member
The input FETs are not safe above 30V or so. You will need a cascode.

It will work fine if you use the cviller F5c boards and use just one output pair. Keep dissipation within the 30W target per device and you will be fine.

400VA sounds about right as a universal power supply.