ESP P3A

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Disabled Account
Joined 2007
if the RCA (or other input socket) has any exposed conductive part then the RCAs must be connected somewhere to chassis.
Leach and many others show this.

To safety earth, you mean. Yes, this is true but the diodes provide that connection. :bomb:

All roads lead to Rome......all loads must have their own unmodulated ground connection paths.

A perfectly acceptable and often used grounding strategy is to use more than one star ground.
 
If I want to change the bias transistors to something that can be mounted on either the output drivers or output transistors heatsink is there anything I should look for? Or can I just slap a bd139 in that posistion and be done with it?

And is it really necessary? Im not planning to increase the bias any higher than the 100mA specified by Rod.
 
Last edited:
The output bias depends principally but not exclusively on the driver device Vbe variations with temperature.
The temperature compensator must as a minimum monitor the driver device temperatures.
Locating the two drivers and the Vbe multiplier on a short thick strip of aluminium does this.
 
Well, found a picture of the real deal. Rod has just placed Q9 close to Q6.

With my current layout I can move a few components and get Q9 close enough to touch the heatsink of Q5. The flat side of the package of BC546 would be up against the heatsink. That should be adequate.

Like this :
 

Attachments

  • P3A-V3.PNG
    P3A-V3.PNG
    121.9 KB · Views: 1,003
Last edited:
Could one use a bd139 for that as well or should it be something else?

I prefer the better thermal tracking of a metal (+insulator) interface e.g. BD139, especially with EF o/p's, but Doug Self shows that the driver of a CFP stage isn't subjected to much temperature variation, so it may not be so necessary.

FWIW, Carlos's DXBlame (EF stage) uses a BD139.

Brian.
 
Use the MJE15034/5 drivers. Don't mount TO-3's on board, connect them with wire - just keep it short. Keep Q1 and Q2 close - ideally, position them so that the flat faces of the TO-92 packages can touch and be bonded together with a wire tie. Keep Q3 close to the LED.

Q4 will get hot - leave space to fit a small heatsink. Keep C4 and C6 as close to their devices as possible - ideally, find 2.5mm pitch parts. Regardless of what most audiophile types will say, good high stability (NPO type) ceramics work very well here.

The Zobel resistor R15 should be a 1W type - I personally fit the Zobel network across the speaker terminals and not on the PCB. This worked well for me.

Try to get the layout as condensed as possible. There's some rather long traces on there you could get shorter.
 
Made some major changes.

Driver transistors will be Tip41/42 or MJE150034/35(have the same pinout)

The bias transistor changed from BC546 to BD139.

Both drivers will now be connected to the heatsink, the same goes for the bias transistor, should prevent thermal runaway.

The 2 x BC546 in the LTP are now facing each other with the flat surface, should give them decent thermal coupling with a zip tie around them.

Moved the LED up against the CSS transistor.

The 2 x 100 pF capacitors are now placed up against their corresponding transistors.

Gnd is now separated into, signal gnd, zobel gnd, +V supply gnd, -V supply gnd.
 

Attachments

  • P3A-V2-SCH.png
    P3A-V2-SCH.png
    59.1 KB · Views: 859
  • P3A-V2-BRD.png
    P3A-V2-BRD.png
    163.5 KB · Views: 855
Hi,
you need to add a trimmer across R5 or R6 to adjust the currents through the LTP to exactly match.
But you have no method of measuring the currents for matching.
Add degeneration resistors to the LTP. You may need them to set up the compensation.
Then you can also use them to measure the currents.
It would be good to allow space/pads for adding degeneration to Q7

C6 is too small. The DC blocking cap, on the input, has F-3dB set to ~1.4Hz. Either increase C6 or decrease C11 to ensure that the NFB has a wider bandwidth than the input filter.

BC546a is a poor selection for your LTP.
b grade will be better suited, but c grade or better is what you should have at your input. BC546c are not easy to source.
 
Last edited:
Hi,
you need to add a trimmer across R5 or R6 to adjust the currents through the LTP to exactly match.
But you have no method of measuring the currents for matching.
Add degeneration resistors to the LTP. You may need them to set up the compensation.
Then you can also use them to measure the currents.

C6 is too small. The DC blocking cap, on the input, has F-3dB set to ~1.4Hz. Either increase C6 or decrease C11 to ensure that the NFB has a wider bandwidth than the input filter.

Well, its a straight copy of Rod Elliotts 3A amplifier, except a few of the transistors have been changed. I do get what you are saying, but dont you think he would have made the schematic with those changes if they were really needed?

Why does the NFB need to have a wider bandwidth than the input filter?
 
Last edited:
BC546a is a poor selection for your LTP.
b grade will be better suited, but c grade or better is what you should have at your input. BC546c are not easy to source.

Dont put too much into the component names, some of them are just chosen for their footprint. I am going to use BC546C. Same goes for the drivers, they are tip30/31 in my schematic but I am going to use something like MJE15034/35 or MJE15030/31, which has the same package/pin out. :)
 
Last edited:
Yes matching will help. The emitter degeneration resistors AndrewT mentioned will help performance but then you are modding the circuit which gets more involved. It works well as is.

Increasing C6 is a good idea. I put 330u here. It's a simple change that does help the sound.

BC546B works well, thats what I used on mine :) If you buy them in a batch (mine were Onsemi samples) theyre typically quite well matched already
 
Yes matching will help. The emitter degeneration resistors AndrewT mentioned will help performance but then you are modding the circuit which gets more involved. It works well as is.

Increasing C6 is a good idea. I put 330u here. It's a simple change that does help the sound.

BC546B works well, thats what I used on mine :) If you buy them in a batch (mine were Onsemi samples) theyre typically quite well matched already

Well, I made room for a 10mm, 5 mm pitch capacitor as C6 so theres plenty of room for a bigger cap. It afterall only has to be a lov voltage type.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.