F5 Turbo Builders Thread

Moderator
Joined 2002
Paid Member
Finally got a build done :D

Have 1.5A bias for Class A output up to about 35W, expect this to drop during summers as the thermistors clamp hard. I am using a thermistor on each device. Higher is not possible within this kind of layout.

All things considered, it sounds great. Biased up first time without any issues. One channel was distorting badly at high levels, traced down to a frayed input wire. As simple and trouble-free as the original.

Props to Mr. Pass for a great amp. I still prefer the sound of the original F5 (I think, as I savaged it to get some parts for this build). I only had an old pic to put up....

Using two banks of 100,000uF for each of the amps post R. The forward sink takes a bit of time to reach temperature, but I still have under 50 degrees on the rear sink (adding a bit of headroom for the heatspreader). I am slowly convinced that my bigger build will turn out quite well, provided my back holds up.

I have since installed all the panels, and added a heat barrier between the caps and the heatspreader. The transformer runs warmer than I expected, even though it's a pretty solid 400VA unit. I may need a totally perforated top panel, but we'll see.
 

Attachments

  • 22.jpg
    22.jpg
    583.7 KB · Views: 401
Moderator
Joined 2002
Paid Member
Yup, I know :D Those are the front sinks, the rear is at the bottom of the pics.

I had planned on mounting the cap boards at right angles and mounting the PCB in the center - but for various third-world logistics reasons that didn't work out and I needed the build finished while the ambient temperatures were low enough to arrive at a reasonable level of bias. I've been harassed about space between the outputs plenty times in the past, but in practice it's not as bad as it seems.

The same sinks, when in one piece were 10" tall and could deal with 80W of bias though a single pair was very stressed out at that level. This version can do just a hair under, the added thermal mass of the 8mm heatspreader helps with keeping temperatures very similar at similar bias levels.

In any case after a few hours run the temperature difference between the sinks is just 5 degrees at the sink center. Initially they run quite cold, but catch up given time.

I also am depending heavily on the fact that the amps will be running flat out for a large portion of time. In which case the temperatures are usually much lower than at idle anyway.
 
Moderator
Joined 2002
Paid Member
Very nice, sangram! True DIY.

Walter

Thanks Walter for the kind words :)

If an F5t has a maximum power of 60W

My specific build is a bit different, I did say that the Class A bias was limited to 35W of the output.

And the words 'flat out' do not refer to operation at 0dB - I don't know what you thought I meant. That would be foolish. 'Flat out' is a colloquial expression, not a technical definition.

They refer to the maximum power level my ears can take at the listening position. Obviously this does not depend on the power output alone, as you know very well. Listening distance, speaker sensitivity and program characteristics all have a role in defining this.

In any case. I have a simplistic way of looking at the bias level and temperatures.

Assume power output for loud listening level is 2W into a reactive load.

My bias is set to 1.5A at 32V rails. This is 96W into the sink.

My sink temperature is 50 degrees at ambient of 20. Give or take, this works out to about 0.3C/W.

Now at an output power level of 2W into 8 ohms the current drops to 1A (500mA/8 ohms) as 500mA is passing through the load.

This means the power into the sink drops by 30%, and the temperatures fall into the mid-40s. Or lukewarm sinks. I probably could have pushed the system harder, but have chosen a bias point that would suit the climate variations in my part of the world. we regularly see ambient temps of over 30C indoors.

Shockingly, my actual temperature readings correspond quite closely with the numbers here.

It's not the perfect or ideal build in your book, I get that. It still works just fine for the application I executed it for. It was to also contain a tube preamp and a USB DAC, but that got shelved quickly as there was not enough space.
 
...........................

My bias is set to 1.5A at 32V rails. This is 96W into the sink.
...........................

Now at an output power level of 2W into 8 ohms the current drops to 1A (500mA/8 ohms) as 500mA is passing through the load.

This means the power into the sink drops by 30%, and the temperatures fall into the mid-40s. Or lukewarm sinks. I probably could have pushed the system harder, but have chosen a bias point that would suit the climate variations in my part of the world. we regularly see ambient temps of over 30C indoors.

Shockingly, my actual temperature readings correspond quite closely with the numbers here......................
I don't know how you managed to measure a drop in temperature delta of around 30%, but the explanation you give is complete fiction.

If your average output power is 2W and the quiescent power was 96W, then the average dissipation is 94W, while average output is ~2W
That will NOT lead to a reduction of 30% in delta T !
 
F5 Turbo build of Friend, playing nicely

My friend got help for soldering these part over PCB. He made his own chassis, heatsink height is 7", footprint is 12"X 24". I have completed wiring for this. Both channel are biased for 0.6A per device and running for 15 hours since last 4 days. Heatsinks are stable and could be touched all the time forever. Sounding nice with ~80W/ch biasing. Output offsets are ~11mV, -3mV after 15 hours continuous running. Complete silent, no hiss no hum. This is my third F5 turbo.

I haven't added MUR3020 diodes on amp board yet but planning to do so.

Big 1KVA transformer with 4 x 24-0 secondaries. Using softstart board from *bay. Ground isolator are from both channels 0 level are hidden below softstart board.
9WSq23.jpg

Power Supply with discrete diodes making 4 bridge rectifiers for 2 channels. Ground is derived at capacitor bank.
udYtNv.jpg

Board wiring-
UTyjJy.jpg

See the signal wires, input is Gotham GAC2 cable with screen grounded at RCA socket -
qnryDc.jpg

Over all beauty missing faceplate yet.
c0WMCi.jpg

Its huge build with >30Kg weight. We need 2 people to move it around.
 
Last edited:
Hi
Did someone had the chance to compare the F5 Turbo with KSA50 (clone) or original.
Reason I ask because I do have a 1KVA 30-0-30VAC transformer, heatsink etc.
Out of the two amp which would be better option to build.
May be someone built both amplifier can share his opinion, experience

Greetings
 
Thank you
I know that! It is great we can get the parts and PC boards readily
(but) I also have a pair stuffed KSA50 PC board at home.
The reason I posted here that question may be someone built both amplifier and able to share his opinion or experience. A lot of KSA was built and F5T..
If the F5T V2 sound better I would sell the KSA stuffed boards and go for the F5T.:)
 
Moderator
Joined 2002
Paid Member
If your average output power is 2W and the quiescent power was 96W, then the average dissipation is 94W, while average output is ~2W

2W/8E requires 500mA of current.

2W/4E requires 700mA of current.

2W/2E requires 1A of current.

Please expand my knowledge by explaining how the amplifier dissipation remains exactly the same under all three conditions above.
 
Hi
Did someone had the chance to compare the F5 Turbo with KSA50 (clone) or original.
Reason I ask because I do have a 1KVA 30-0-30VAC transformer, heatsink etc.
Out of the two amp which would be better option to build.
May be someone built both amplifier can share his opinion, experience

Greetings

I compared back to back he F5 (standard) with a totally refitted KSA200. The F5 was quieter, electrically, and had better definition in the bass. The Krell was slightly loose and overblown in the bass giving the subjective feeling of being a little slower. This was into Ns1000's so things might change with a tougher load.

I'd take the F5 in a heartbeat.
 
2W/8E requires 500mA of current.

2W/4E requires 700mA of current.

2W/2E requires 1A of current.

Please expand my knowledge by explaining how the amplifier dissipation remains exactly the same under all three conditions above.
Do you want to listen, or would I be wasting my efforts?
At the moment and for my last few posts you are clearly not bothering to read the content.
example:
I posted
quiescent power was 96W, then the average dissipation is 94W
and you replied with
the amplifier dissipation remains exactly the same
 
Last edited:
I can't really explain what's happening, but by sangram's explanation 6W into 8 Ohms would lead to 100% efficiency, zero dissipation in the devices.

I think one has to look at the voltages. These 500 mA into 8 Ohms only create a 4V swing. Where's the rest of the rail going?

It's still confusing me a lot, all this class A push-pull thing.