[f] AMB on Facebook

Initial set up issue β24 board

beta24

Initial set up issue β24 board

Postby Pieterh » November 29th, 2017, 11:47 am

Dear Ti, I completed the initital set up of the first B24 succesfully. However, with the second one I am confronted with an issue.

I set R7 and R9 at 4.0 volts with VR1, but then R8 and R9 are 4.1 volts (0.1V higher).
I set the voltage across R11 and R13 at about 129mV with VR4.
I can set the voltage across R45 and R46 correctly both at about 40mV. But if I set R47 at about 40mV, the voltage across R48 is always about 1.5 mV lower. With the first board, i didn't have these differences.
The voltage across the 22R-10W resistor on the positive side is 8.6V and on the negative side 9.6V, about 1V difference.
At this point of the test cylcus, the offset across the Output + and output- is about 350 mV, and then if I adjust VR4, which I have to adjust quite a lot to get a 0 mV offset , the voltage across R45 and R46 goes down to only a 1-2 mV.

I hope you have some clue about what happens?
Pieterh
 
Posts: 132
Likes: 93 posts
Liked in: 54 posts
Joined: December 29th, 2015, 12:50 am
Location: Netherlands
Country: Netherlands (nl)

Re: Initial set up issue β24 board

Postby amb » November 29th, 2017, 12:26 pm

You can think of the amp circuit as a quadcopter, when all four engines are working in equilibium, then the craft hangs perfectly in the air in balance. The β24 circuit has four quadrants that needs to similarly hang in balance for it to work properly. Unfortunately, due to parts tolerances that could only be achieved by introducing slight variations in order to bring things back to balance. There are occasions where the adjusting VR4 too far could cause one side of the amp to lose bias, and that appears to be your case. But you need to go that far in order to being the differential output DC offset down to near zero.

When this happens, my advice is to start over with the initial setup procedure (resetting all trimpot to their initial positions as documented), but this time set VR1 until the voltage across R7 is 4.5V instead of 4V. The voltages across R8, R9 and R10 should also go up accordingly. Now continue with the rest of the setup procedure and see if you could get the amp the behave well. If a similar problem still occurs, you can repeat again and try 5V across R7.

Pieterh likes this post.
User avatar
amb
Site Admin
 
Posts: 10380
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Initial set up issue β24 board

Postby Pieterh » December 3rd, 2017, 12:33 am

No succes, unfortunately. I set 4.5V over R7 and then VR1 was a maximum position (so, i can't try 5V), voltages across R8, R9 and R10 went up accordingly but were all about 0.1V higher. The voltage across R11 and R13 were 148 mV. then the voltages across R45 and R46 were about 40mV again. and R47 was 40mV and R48 was again on average about 1,5 mV lower, so about 38,5mV. Voltages across 22R10W were about the same about 8.9V. The offset was 330 mV (20-30 mV lower than last time with 4V across R7) and after adusting VR4 and offset = 0, the voltages across R45 and R46 went down to 5.9mV (last time with 4V / R7 this was 1-2 mV). The voltages across R47 and R48 went up to 58 and 57 mV. R11 = 37mV, R13 was 272 mV. At this point I measured the voltage across R7 and R9 again and this was 4.7V and across R8 and R10 = 4.8V.

I did observe that the Q9 and Q13, Q10 and Q14, Q11 and Q15, Q12 and Q16 transitors became quite hot during the set up. After adjusting the offset to zero, Q9 and Q13 and Q10 and Q14 cooled down. The other transistors Q11 and Q15, Q12 and Q16 kept being hot.

I did measure HFE for the BC550 and BC 560 parts. I give the values below and also in the attachment within the scheme, it may be usefull info.
B24_incl_hfe_550_560_v_small.zip
However, these approximate values were also installed on the other board which set up was fine. These are approximate numbers as I put the transistors in 'groups' that had a certain range.
Q5, Q7 = 750
Q6, Q8 = 520
Q9, Q13 = 500
Q11, Q15 = 500
Q10 =685, Q14 = 700
Q12 = 685, Q16 = 700
Q17 = 750
Q18 = 520
Q19 = 750
Q20 = 520
Q45 = 650
Q47 = 680
Q49, Q51 = 500
Q46 = 500
Q48 = 480
Q50 = 650
Q52 = 680
You do not have the required permissions to view the files attached to this post.
Pieterh
 
Posts: 132
Likes: 93 posts
Liked in: 54 posts
Joined: December 29th, 2015, 12:50 am
Location: Netherlands
Country: Netherlands (nl)

Re: Initial set up issue β24 board

Postby amb » December 3rd, 2017, 9:13 pm

From the + side to the - side your BJTs are well matched. That should be enough to make the adjustments work, but since it doesn't, try replacing R7-R10 with 2.21KΩ resistors. Reset all trimpots again and repeat the initial setup and see how it behaves.
User avatar
amb
Site Admin
 
Posts: 10380
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Initial set up issue β24 board

Postby Pieterh » December 8th, 2017, 8:50 am

Hi Ti, I have some additional information that I want to share with you before changing the R7-R10 as advised above (I got the resistors today so I can do it next if needed). I had read through the forum yesterday about bias issues and found info on long wiring and adjusting vr2, vr3 and vr4 more or less together to find the best offset while stil keeping the bias about 40mV across R45-R48. So, today I got rid of too much wiring. But this didn't solve my issue. So I resetted all trimpots and started over again. So I had VR1 at max, then adjusted VR2 to have 40mV bias, adjusted VR3 to ahve 40mV bias, then adjusted the offset with VR4 from 330 to 150mV. So,now the bias across R45-R46 went down to about 20, so I adjusted VR2 clockwise to get 40mV again, I adjusted VR3 anti-clockwise again to reduce the higher bias to about 40 mV again. I repeated this procedure with correcting the offset downwards, then VR2 clockwise and then VR3 anti-clockwise. VR2 was now in max clockwise position. The best offset I got in this way was between +20-50mV. All biases were about 40mV (except R48 which is always 1,5mV lower). The point is that the R11 value was about 65 mV and R13 253 mV. So very much out of balance. the same was true for R12 (60mV) and R14 (250mV). Q12, Q16, Q15 and Q11 were to hot to keep my finger on and my question is about the cause of that, any idea? The Q10, Q14, Q13 and Q9 had normal temp. Do you have any idea what is causing this unbalance and hot transitors? Could it be that something is wrong with Q19 or Q20 or with one of the current limiting diodes somewhere?

Measured values:
R7 = 4.8V
R9 = 5.0V
R8 = 4.9V
R10=5.0V.
R11= 65Mv
R13= 253mV
R12=60mV
R14=250mV
22R10W = 8.9 (- side) and 9.0 (+ side) volts.
Pieterh
 
Posts: 132
Likes: 93 posts
Liked in: 54 posts
Joined: December 29th, 2015, 12:50 am
Location: Netherlands
Country: Netherlands (nl)

Re: Initial set up issue β24 board

Postby amb » December 8th, 2017, 2:00 pm

The lopsided voltages across R11 and R13 is the cause of the Q11, Q12, Q15 and Q16 to run hotter. Rather than changing the R7-R10 resistors, try one of the following (not both):
- Swap the transistors between Q9 and Q11
- Swap the transistors between Q10 and Q12

Then reset ALL trimpots to their initial positions and re-do the initial setup. Let me know how this works out for you.

Pieterh likes this post.
User avatar
amb
Site Admin
 
Posts: 10380
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Initial set up issue β24 board

Postby Pieterh » December 9th, 2017, 3:53 am

I swapped Q12 against Q10. I choose that couple out of your two options as I think that the unbalance orginates from the section that drives Q36 and Q40. It is just a thought as the voltage across R48 is always 1,5mV lower than the voltage across R47, and I did not see such differences with the other 0.2 ohms resistors (R45 and R46).

Ok, I did not see any significant changes in measured values. I tried only VR1 at 4 Volts across R7. Offset became 330mV. Then adjusting this back to around 60mV with VR4 while tuning VR2 and VR3 to keep about 40mV bias at both sides. Then the R11 (61mV) and R13 (210mV) were again out of balance and the Q11,Q15, Q16 and Q12 became hot again. 22R10W were 8.6 V and 8.7V.

What do you think?
Pieterh
 
Posts: 132
Likes: 93 posts
Liked in: 54 posts
Joined: December 29th, 2015, 12:50 am
Location: Netherlands
Country: Netherlands (nl)

Re: Initial set up issue β24 board

Postby amb » December 9th, 2017, 9:00 am

Ok, then let’s work inwards. Try doing one of the following (not both):
- Swap Q1 and Q3
- Swap Q2 and Q4

Reset all trimpots and do the initial setup again.

As I previously used the quadcopter analogy, since parts have tolerances, when VR4 is in the center of its adjustment range the amp is not in equilibrium, but tweaking VR4 too much causes one side to lose bias. So what we are attempting now is to swap parts from side to side to try to make them less unbalanced. If we succeed then it shouldn’t require as much adjustment on VR4 to achieve that equilibrium.
User avatar
amb
Site Admin
 
Posts: 10380
Likes: 175 posts
Liked in: 482 posts
Joined: March 29th, 2010, 9:49 pm
Location: Sunnyvale, CA
Country: United States (us)

Re: Initial set up issue β24 board

Postby Pieterh » December 10th, 2017, 12:59 pm

I swapped Q1 with Q3. No changes. R13 at 230mV and R11 at 50 mV with an offset of 60mV. R7 = 4.3V. I noticed that the offset is already high from the beginning of the setup procedure (measured after voltages across R11 and R13 were made equal) so without turning up any bias with VR2 and VR3.
Pieterh
 
Posts: 132
Likes: 93 posts
Liked in: 54 posts
Joined: December 29th, 2015, 12:50 am
Location: Netherlands
Country: Netherlands (nl)

Re: Initial set up issue β24 board

Postby mpa2 » December 10th, 2017, 3:00 pm

I would check every resistor first. Wrong parts and bad joints, the "human factor" :) Since you have one working board, measure every resistor value with your DMM and compare it with the "good" board (reset the trimpots first). And check for cold joints, esp. ground pins, they are often hard to solder properly.
mpa2
 
Posts: 77
Likes: 2 posts
Liked in: 12 posts
Joined: March 6th, 2017, 10:21 am
Location: Germany
Country: Germany (de)

Next

Return to β24 fully-differential power amplifier

Who is online

Users browsing this forum: No registered users and 0 guests

Quick Links

AMB audio DIY resources
AMB audio shop

cron