New Amplifier - ULD Extreme

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
M

mightydub

Member
Joined 2006
AndrewT said:
Hi,
my experiments with 4 NJL diodes and 5 NJL diodes to temperature compensate the output stage bias current have proved unsuccessful.
I tried 4 diodes + resistor at 7mA and at 1.5mA, neither gave anything close to good temp comp. Always goes up with increasing temp.

I tried 5diodes + resistor @ 1.5mA. Bias Voltage far too high for a two stage EF.
Added a bypass resistor across one diode. Had to reduce this resistor so much that the 5th diode was passing <0.5mA (Vf~300mV).
Temp compensation was still poor but better than 4diodes.

I am now in the process of modifying the PCB/layout to incorporate a three stage EF with 5 NJL diodes and a series resistor.

Has anyone else managed to get good temp comp for the NJLs with either 4 or 5 NJL diodes? How did you do it? How well did it perform?
Click to expand...

I have been looking biasing and temp comp with TTraks in detail in simulation - see the thread I started: http://www.diyaudio.com/forums/showthread.php?t=151357

For this output stage, I found (in simulation) that you can get very good temp comp using 4 TTrak diodes with a resistor bypassing the whole string. You have to adjust the bypass resistor to get the right slope, this of course depends on the VAS current. Then adjust the resistor between the driver emitters to adjust the idle current. This assumes the drivers are mounted on the heatsink with the outputs so they track thermally.

With 7mA VAS current, bypass resistor = 470R and resistor between driver emitters = 360R is between 108 and 109mA between 20C and 100C. Beyond 120C though it starts to increase sharply so thermal protection should kick in at 100C - that would be prudent anyway.

Again, this is simulated results, your mileage may vary. Maybe somebody can try this out?
 

Attachments

  • ULD_ops_thermal.JPG
    ULD_ops_thermal.JPG
    54.9 KB · Views: 722
A

AndrewT

R.I.P.
Joined 2004
mightydub said:
With 7mA VAS current, bypass resistor = 470R and resistor between driver emitters = 360R is between 108 and 109mA between 20C and 100C. Beyond 120C though it starts to increase sharply so thermal protection should kick in at 100C - that would be prudent anyway.

Again, this is simulated results, your mileage may vary. Maybe somebody can try this out?
Click to expand...
your simulation seems very similar to what I said did not compensate.
 
A

AndrewT

R.I.P.
Joined 2004
M

mightydub

Member
Joined 2006
AndrewT said:
read
http://www.diyaudio.com/forums/showpost.php?p=1896550&postcount=321

Used a parallel resistor to adjust the diode current.
Used a series resistor to increase the bias voltage to turn on the output bias.
It fails to compensate for output temperature.

If your simulation appears to turn on the outputs with just 4diodes at a reduced current then your model is wrong.
Click to expand...

Andrew, I did read your post very carefully before my first post, however it wasn't exactly clear how you had the bias configured. If it is the same as the schematic I posted - well, that's why a picture is worth a 1000 words.

I agree that the simulation will only be as accurate as the models, I used the models direct from OnSemi for all components. If you, or someone else can measure the diode forward voltage and let me know the current and temperature I can compare that with the simulation. (I don't have any TTrak parts on hand.) AndyC has posted some modified models for the diodes, I can have a look at the behavior with those as well.
 
C

Carl_Huff

Member
Joined 2004
kadproducts.com
Attention all,

Hugh Dean has been busy working with Peter Doyle to debug his amplifier and they have found two layout mistakes! Both undoubtedly slipped in at the same time and are located in the same area of the PCB. The good news for those of you that have fabbed PCBs the fix is an easy one. By the time you have read this I will have updated the layout and gerbers that are online for download and post the URL again.

Error #1 before the fix
 

Attachments

  • ERROR_01.jpg
    ERROR_01.jpg
    10 KB · Views: 802
Last edited:
AKSA

AKSA

Member
Joined 2001
aksaonline.com
Thanks Carl!!

No cigar yet. :(

Peter still has to come back in ten days for instalment two, where we fix the second mistake and actually trial the amp on the test bench.

These errors might require new boards as they are tricky to fix on the existing ones, but then, there are only three boards in existence, no big deal.

In the meantime Carl will work on the existing pcb and put up new gerbers for those who would like to build their own. ;)

Hang off until we get it going and playing music....... :cool:

Cheers,

Hugh
 
T

tryonziess

Member
Joined 2007
I love this place. All of these highly skilled folks lending a hand for FREE.
When you start your project you know someone better than you has been there before. Good work. When you get it all tested I think I will try one of these. I think there are a few board houses still making small board orders for around 33.00.
Tad
 
R

rumors

Member
Joined 2003
AKSA said:
Jeremy,

Another interesting fact is that the higher this resistor, the lower can be set the quiescent current. The math, based on crossover considerations, indicates that 26mV (derived from the Boltzmann constant) should be dropped across this resistor regardless, so 0.22R is passing 120mA and 0.1R SHOULD strictly be passing 260mA. It isn't in this design, so immediately there is one compromise....

Hugh
Click to expand...


Based on this observation if raise the RE 0.22 to 100-ohm and connected in parallel to these capacitors 10000uF 10V?
Dynamic class A ?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom