How to build the F5

Yes, I guess it was R11 and R12 smoking. Just goes to show how robust the output stage is.

I don't think you could run it for very long even with a higher wattage R11 and R12. Do bear in mind that Tj would still have been close to ambient (11 degrees), after a very short time Tj would rise rapidly and the output devices would probably self destruct.
 
Last edited:
Where would you connect the Speaker 0V ?

Just after a little bit of advice here.

Here you can see my F5 nearing completion. The burning question is where would you connect the O/P 0V return ?

My best bet would be at point A where the transformer 0V meets in the middle of the PSU. The orange boxes indicate the PCB layout of the PSU, yes its just huge areas of copper.

The PSU is a CCRCCCCC.

22000uF, 4700uF, 0R136, 4700uF (x5)
 

Attachments

  • DSIR9631.JPG
    DSIR9631.JPG
    222.8 KB · Views: 420
Last edited:
Hi, this is for Kate, yes?

This may seem a bit pedantic and impractical at this late stage of your assembly but ....

As you have it, there's one large power supply for both channels, and no problems that I can see - however, if you were to move the diode bridge (an IXYS, I hope?) and the first power capacitors, right beside the transformer, this would shorten the secondary wires and hence the radiated 'noise', current spikes, etc

As that transformer takes up a lot of floor space in the middle, you can use a right angle bracket (strong Aluminium is good) to mount it on it's side, right up front (or with a cross brace) across the front of the case - then position the block bridge and the first filter caps right next next to this and creates more space to rotate the supply caps (pcb mounted ones) by 90* in the centre so the raw current spikes from the diodes are going directly into the pcb near the transformer, diode ripple cap end and the output rail wires feeding to amp boards are at the rear of the case (going sidewards, if that makes sense!)

Now, about the centre 0 volt point - the raw currents of the transformer secondary windings go into the diodes then into the filter caps are what you would call a 'local current loop' and so the centre tap of the transformer is connected directly there -

However, you have complicated things a bit by added another pair of caps in parallel (first ones on the pcb) and this also has it's midpoint 0 volt connected back to the centre point of the first ripple caps via the end terminals of the pcb - this is the 'raw supply local 'star' 0volt point (it may seem to be a bit 'over the top' but makes a huge difference to how quiet the amp sounds)

The other end of the pcb is normally used as the main, central 0 volt point of the whole amp, including the connection via the thermistor to the chassis ground point next to the IEC socket, speaker return from terminals, amp pcb, etc

Yes, that length of pcb trace between the first 2 caps on the pcb and the other end of the 8 filter caps does make all the difference worthwhile the trouble.

All this requires is another set of holes in the base plate of the case and no cosmetic alterations.

Now, about that - I would move both the input RCA phono sockets, and the output speaker terminals, right out towards the sides of the back plate closer to the amp pcbs and move the input AC socket, (fuse & power switch?) to the centre of the back panel and run the ac wires to the front transformer along the bottom plate - I twisted all 3 wires and added a mesh outer cover connected to the metal base as an ac shield, but was surprised at the small difference this made, so wouldn't bother about this.

... my 2 cents - hope it's of use.
 
Thank you James for all the pointers.

Unfortunately I'm not going to be drilling any more holes in the chassis so the connectors to the outside world will remain as they are.

This amplifier is part of a tri-amp system, it will be driving the tweeters in my speakers. As such, any miniscule 50/100Hz hum is largely irrelevant.

The mid is being driven by a Pass Aleph 4 which has a noise floor lower than a snakes belly, dual mono construction with HUGE power supply.

The sub 400Hz is being driven by yet another 150W Mos-Fet amp, again with a pretty massive PSU.

Hopefully each amplifier will be working at its optimum - time will tell.
 

6L6

Moderator
Joined 2010
Paid Member
Just to clarify, you say it's fluctuating between +/- 6mV? 0.006V?

That's so low as to not matter, and anyway, apply a signal and it will fluctuate much more than that.

Essentially you are down in the noise that can only be fixed by lead dressing and the like. I'm sure Andrew will suggest twisting your leads and such. :D :D :D

Your build looks fantastic, I would button it up and listen to it for a week and then re-check the bias and be done.

Congratulations!!
 

6L6

Moderator
Joined 2010
Paid Member
Is it normal to have a bigger offset when first turned ON.

Yes, very much so.

The little thermistors near the power Mosfets are part of a temperature compensation circuit for the bias, and they try to keep the bias stable as things heat up, but they can only do so much. That's why Nelson suggests making the final adjustments to the bias once it's been on at least 1/2 hour.
 
Board Images

Here are a couple shots of my board, as well lit as I can manage in my dungeon.

I believe my Fets are correctly positioned and I don't see any shorts.

I have found that Offset rises faster than Bias until .3-.35Volts (Bias). After that point DC offset levels off around .4volts and I can continue to increase bias.

Your help is very much appreciated.

Happy Turkey Day to those of you in the states.
 

Attachments

  • F5 Left.jpg
    F5 Left.jpg
    32.4 KB · Views: 360
  • F5 Middle Board.jpg
    F5 Middle Board.jpg
    27 KB · Views: 270
  • F5 RIght Side.jpg
    F5 RIght Side.jpg
    34.5 KB · Views: 191
  • F5 Back.jpg
    F5 Back.jpg
    29.4 KB · Views: 187