F5 power amplifier

So everything is OK - you just can't expect the 25W amp to drive the speaker to same SPL as a 100W amp does.

Thanks, Juma!

So I think I was right with my assumption...
The sound was perfect and clear, only the bass wasn't as loud as on the 100W amps.
And that seems to be normal- it's only a quarter of the power.... :)

If I put one MosFet parallel to the other, and raise the 590mV, maybe it would be a little bit better?
The PS ist well dimensioned (500VA and enough caps per mono side), so that shouldn't be the problem?

regards,
Matthias
 
i have dual outputs and bias at 1.7A(400mV pr fet. so 800mV pr rail)
and that work very well on my scan-speak 3Way with dual 8" in the bottom.
and only 225VA and 74.000uF pr ch:) dissipation is about 86W at the sinks.


hi audiosan!

can you tell me, if its enough to put the output-mosfets parallel to the existing ones, without doubling the input parts or ztx devices?

THANK YOU!

regards,
matthias
 
here is the finnished board. without the fets.
 

Attachments

  • DSC00329.JPG
    DSC00329.JPG
    916.9 KB · Views: 526
67 year-old newbie begs for a little help !

Hi Gang: I'm a semi-retired recording engineer, pretty well-known around the old US of A. I've built a lot of stuff over the last 59 years, so have good construction knowledge and ability.

Problem: I've just joined the forum, and two nights ago I discovered the F5 build posts. I'm totally fascinated, and would like to build a couple of monoblock F5's.

25 watts is plenty, as I intend to use them with Wayne Parham's Pi-4's, which are very sensitive at 98 db or so.

The problem again: The info chain and posts have been going on for several years now, and it appears that build info is scattered among thousands of posts.

Is there someplace where the latest info, BOM and vendors can be found? The idea of culling several thousand posts to try to figure this out is a bit daunting, and I'll bet some of you have nailed this info and made it available easily. I'll order boards (which ones?) from the Store and sure will appreciate getting pointed in the right direction for info for a standard F5 build, BOM, and parts sources.

I sure hope that I'm not asking for too much, and many, MANY thanks for the help!

Jeff:)
 
hello. a had the same problem:) the F5 interest came to late:)
well. i can help you with a BOM.
however. what boards are the best for you, depends on your heatsinks. (so you can spread the heat evenly) but the store's boards are wery good. at least for sinks thats max 300mm. over that, or with 2 sinks conected together. a wider space between the fets will be good.
here is a BOM:)
 

Attachments

  • bom F5 v2.pdf
    50.9 KB · Views: 228
Is there someplace where the latest info, BOM and vendors can be found? The idea of culling several thousand posts to try to figure this out is a bit daunting, and I'll bet some of you have nailed this info and made it available easily. I'll order boards (which ones?) from the Store and sure will appreciate getting pointed in the right direction for info for a standard F5 build, BOM, and parts sources.

I sure hope that I'm not asking for too much, and many, MANY thanks for the help!

Jeff:)

Really you just need to read the F5 manual, then read the F5-Turbo manual (First Watt website).
I built a couple of F5s and decided to leave out the current limiting section. I have the F5c boards and some others. The F5c are great as you can build a F5-Turbo with 2 pairs of outputs per channel (matched) and use a higher volt rail (32 volts) if you want and cascode the input jfets (Nelson said the turbo sounded more relaxed). Or just build a regular F5 with the same boards. Lots of choices from one set of boards. The power supply boards from the store are also good.
Most stuff you can get from Mouser or Digikey. Heat sinks from Heatsinks USA or complete chassis with heatsinks from eBay sellers (soon from the store). There are several DIYAUDIO members selling boards an outputs and various parts. Buzzford, Zhoufang, Jackinnj, Peter Daniel, check the swap meet forum section.
I have some 2SK170BL and 2SJ74BL if you need them. PM me. 2SJ74BL are getting hard to find.
Hope this helps.

Rush
 

6L6

Moderator
Joined 2010
Paid Member
Last edited:
Problem with paralleling outputs

I have been happily working with an F5 and a single MOSFET output for a year (2SK175/2SJ55) and it worked well.
Well, I decided this week to start to quadruple the output, as I had installed four pairs per side on the heatsink already.

I ran into problems

audio F5 2012  4 65.jpg
Top output bottom input (the HF oscillations are reflected to the input even!); the oscillations can be seen first on the upgoing sine from bottom.

audio F5 2012  2 64.jpg

  1. Instable - there was instability as I increased the bias in the right channel: severe oscillations beyond a point of about 0,8 Amp total in right Ch. Left can get more and I can set the bias there to about 1,5 amp total.
  2. ringing with a capitative load of 100 nF; a bad issue because I have Quad 57 electrostats.

I noted that when I did a sine on 8 ohm with high volume the instability could start to appear (first on the negative rail).
This issue can be heard on the speakers. There is popping sound and distortion as the volume goes up.

The second issue of ringiong I did not have before in my single version F5, but with an input RC (1k2/1 nF=78KHz) the ringing is not triggered and disappears.

What can be the problem? The N/P sets were selected from a large batch; I have 0,47 ohm source resistances.
Probable cause:
I did have an open connection to -Vb (-23V) when I started (ampere meter not connected) :xeye: and dialed in the positive side this led to a gate voltage of 23 volt on the 2SK175's (they can stand 20 volts Vgs).​

The 2SK175's do work though - on the test bench the right channel gives a flat output to >600 kHz without an input filter and it works well up to clipping. At higher volume, the DC shifts. This all at a low bias. In the room, the R channel exhibits clicks and pops (probably as the instability occurs).

Is this the typical behavior of a 'half-dead' MOSFET?

The ringing is of course independent of the half-dead devices, if I am right.

happy with any light that is shed.
 
Does each FET have its own gate stopper resistor? Perhaps you should show the exact schematic and a photo of the layout.

I have been happily working with an F5 and a single MOSFET output for a year (2SK175/2SJ55) and it worked well.
Well, I decided this week to start to quadruple the output, as I had installed four pairs per side on the heatsink already.

I ran into problems

View attachment 299090
Top output bottom input (the HF oscillations are reflected to the input even!); the oscillations can be seen first on the upgoing sine from bottom.

View attachment 299091

  1. Instable - there was instability as I increased the bias in the right channel: severe oscillations beyond a point of about 0,8 Amp total in right Ch. Left can get more and I can set the bias there to about 1,5 amp total.
  2. ringing with a capitative load of 100 nF; a bad issue because I have Quad 57 electrostats.

I noted that when I did a sine on 8 ohm with high volume the instability could start to appear (first on the negative rail).
This issue can be heard on the speakers. There is popping sound and distortion as the volume goes up.

The second issue of ringiong I did not have before in my single version F5, but with an input RC (1k2/1 nF=78KHz) the ringing is not triggered and disappears.

What can be the problem? The N/P sets were selected from a large batch; I have 0,47 ohm source resistances.
Probable cause:
I did have an open connection to -Vb (-23V) when I started (ampere meter not connected) :xeye: and dialed in the positive side this led to a gate voltage of 23 volt on the 2SK175's (they can stand 20 volts Vgs).​
The 2SK175's do work though - on the test bench the right channel gives a flat output to >600 kHz without an input filter and it works well up to clipping. At higher volume, the DC shifts. This all at a low bias. In the room, the R channel exhibits clicks and pops (probably as the instability occurs).

Is this the typical behavior of a 'half-dead' MOSFET?

The ringing is of course independent of the half-dead devices, if I am right.

happy with any light that is shed.
 
Or possibly reduce the number of power fet pairs receiving the signal - ie, lift the gate stoppers on 2 of the Fet pairs and see if the input jfets are happier with the load - may not be the answer but quite simple to try, and if not, easy to replace.

There was some mention a long time back about a small Gate/souce cap on the power fets to solve oscillation problems of higher capacitive loads, I think ....
 
Does each FET have its own gate stopper resistor? Perhaps you should show the exact schematic and a photo of the layout.

You mean "each JFET" gatestopper -- not absolutely necessary, but it can be done with slight improvement of measurements.

Of course, the gatestopper for the MOSFETs has to be as near the gate as possible.

Using multiple output devices? -- the input stage might be running out of steam due to the paralleled input capacitance of the output devices.