BAN Build Thread

Support for JLM Audio Kits

Moderator: Joe Malone

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BAN Build Thread

Post by Joe Malone »

mike wrote:So I just finally finished my first channel and it is alive and working! I am getting normal sounding audio output when I plug a mic in and it seems to be behaving as any mic preamp should.

I am having trouble with the "alignment" process though. I initially tried to do it without my IC1/IC2 plugged in. At that stage I was getting 48.1 V at TP1, then with turning the trimmer next to that, I was able to get TP3 to 24.5 V. Turning the trimmer further (no matter how many times) would not allow me to go any further down on TP3 voltage. Next I was able to turn the other trimmer enough to get the DC voltage across the volume pot down to 5 mV.
With OPA2604AP set the trim pot to centre postion (12.5 turns) and leave it there no need to trim at all.
A further question regarding these trimmers: Are they infinite turn? I feel like I can turn them pretty much forever though maybe that's not actually true.
They will click over and over when they reach the end of the 25 turns.
I am also having trouble determining how I should ground the chassis itself. Currently, the chassis conducts well with itself (ie. from a screw on one face to another screw on another face), but there is no conductance between the chassis and the black ground terminal of the DC power jack. I tried to resolve this by grinding off a bit of powder from the bottom panel with my dremel and soldering a ground cable from the black ground terminal of the DC power jack to it, but it does not stick well to the metal of the chassis, and if I remotely bump it this lead will come free. I am debating drilling a hole and putting a bolt/nut through the bottom of the unit that I can crimp a ground wire to. I just find it peculiar though that for an otherwise comprehensive project the grounding solution for the chassis is missing. Or am I missing something?
We run a wire from the 0v dc connector to the back panel of our cases and then on the bottom panel we remove the powdercoat under the countersunk screw heads with a countersink so all panels connect.
If you bought a BA4 or BA2 rack kit this is already done. You only need to get 0vdc connector wire to one transformer bolt hole on the back panel.
Joe :-)
JLM Audio
Capturing Audio without Injury

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Hey Joe,

My kit is the BAN with no transformers and two OPA2604's. Can you please explain what these two trimmers are doing so I can better understand the need to get them right?
Last edited by mike on Fri Nov 21, 2014 12:51 pm, edited 8 times in total.

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Another thing: With the gain of the preamp to max and no mic attached (just a loose XLR cable), I'm getting some pretty good hum which is reading at -51 dB in my DAW. Is this normal?

2 MB WAV recording:
http://www.filedropper.com/nomiclidlooselyon

The chassis is currently still grounded via the lead I soldered to the bottom panel. I will switch it to one of the screws at the back later, though I don't expect that to change anything. All the ground screw terminals on the XLR board, preamp board, DC jack, and every raw chassis point I've tested test continuous. So it's clearly grounded.

I want to make sure everything is within spec. Noise floor is important to me and one of the reasons I build this.

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BAN Build Thread

Post by Joe Malone »

mike wrote:Another thing: With the gain of the preamp to max and no mic attached (just a loose XLR cable), I'm getting some pretty good hum which is reading at -51 dB in my DAW. Is this normal?

2 MB WAV recording:
http://www.filedropper.com/nomiclidlooselyon

The chassis is currently still grounded via the lead I soldered to the bottom panel. I will switch it to one of the screws at the back later, though I don't expect that to change anything. All the ground screw terminals on the XLR board, preamp board, DC jack, and every raw chassis point I've tested test continuous. So it's clearly grounded.

I want to make sure everything is within spec. Noise floor is important to me and one of the reasons I build this.
if you get -51db with the mic pre at its full 66dB of gain would indicate a open noise floor of -117dB. The mic pre has a high input impedance so cannot be left open like you have tested.

Best test is nothing plugged into the mic pre with the pad switch on (as this will give a 120ohm input load) or no pad switch and load mic input with 51R between pin 2 and pin 3. And see what level you see on your daw meter then?

Is this in one of our BA2 or BA4 cases ? Is it using one of our 48v 520mA power supplies?
Joe :-)
JLM Audio
Capturing Audio without Injury

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

It's a BAN4 project in the case you sent to me, with the 48V 0.52 amp 25W max output adapter you provided. So far I've only mounted one card.

With no input xlr cable, and the pad switched on, the noise floor idles around -50-60 dB, but it constantly fluctuates up and down by around 5 db in either direction. Periodically every 5 or 10 seconds it spikes up as high as -32 or -28 dB.

I'm still wondering: what is my target voltage at TP3? When I set the 10k trimmer half way it's reading ~40 V. It's reading around 34 V at the lowest with the trimmer turned all the way. What does this signify and shouldn't it be 24V (1/2 voltage)?
Last edited by mike on Fri Nov 21, 2014 2:13 pm, edited 7 times in total.

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

I added 0.1 uF bypass capacitors from Pins 4 to 8 on both opamps with no improvement.
Last edited by mike on Mon Nov 24, 2014 1:08 pm, edited 2 times in total.

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Here are some pics of the board as it is. The only "mods" are now that I am using a SSM2212 mounted on a SOIC8 to DIP8 board via the underside (which seem to be passing sound and should be a straight substitution so I presume they're working fine). And I now added 0.1 uF bypass capacitors to the opamps which hasn't helped.

What can I do next? I've inspected and done continuity testing for close solder pads to make sure I don't have any tiny shorts and none have come up. I still can't explain why the lowest I can get is 34V at TP3 :|

http://postimg.org/gallery/jqo9kao0/9a6e1a4b/

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Here is a screencap of what the waveform looks like when the noise bumps up to around -32 dB every so often.
You do not have the required permissions to view the files attached to this post.
Last edited by mike on Mon Nov 24, 2014 12:47 pm, edited 1 time in total.

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Further testing: I just completed a second channel. This one is measuring normal ~25 V at TP3. But it's still got the fluctuating noise floor.

I'm starting wonder if I have two problems.

1) First channel measuring abnormal ~38 V at TP3 when trimmer turned half way (for unknown reasons)

2) Faulty power supply

The power supply is the only thing in common between the two channels I've built, so that must be the source of the high/fluctuating noise floor. Any way to prove this? I don't have a spare 48V AC-DC adapter lying around to test with.

Both channels were built identically and sound normal besides the noise floor issue.

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BAN Build Thread

Post by Joe Malone »

HI
Do not add 0.1 across the opamp power pins as this will increase high freq SMPS noise by injecting SMPS noise into the half rail in the BAN.

Does the BAN case have 0ohms to 0vdc on all panels of the case?

Are you doing tests with BAN case fully screwed together with lid and everything and screwed into a rack?

Is your A/D D/A interface running at pro level +18dBu or more in and out?

What is the interface?

Let me know the voltages on all channels at every opamp pin for both opamps and SSM2212 pin so I can check this in the simulator how the SSM2122 performs?

Have you tried the standard BC transistors in one BAN to see if it works fine?

Also email me some large well lit clear in focus photos of the top and bottom of the BAN PCBs.

There maybe something wrong with the supply or it is has to small a load on it. These SMPS get quieter with more load. But 2 or more channels should be fine.

Please answer all questions and send me the photos so I can work out your problem.
Joe :-)
JLM Audio
Capturing Audio without Injury

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Joe Malone wrote:Do not add 0.1 across the opamp power pins as this will increase high freq SMPS noise by injecting SMPS noise into the half rail in the BAN.
Okay. Thanks. I've taken them off. It was a problem before I added those though anyway.
Does the BAN case have 0ohms to 0vdc on all panels of the case?
I'm reading 1 ohm from each panel to the ground tag of the AC jack.
Are you doing tests with BAN case fully screwed together with lid and everything and screwed into a rack?
No. I've been leaving it on the ground with the top open. I just put the top on though and there is no difference in any issues.
Is your A/D D/A interface running at pro level +18dBu or more in and out? What is the interface?
It's an Audiofire 12 and yes it's running at balanced line levels.
Let me know the voltages on all channels at every opamp pin for both opamps and SSM2212 pin so I can check this in the simulator how the SSM2122 performs?
CHANNEL ONE (reading ~38 V at TP3 for unknown reasons)

IC1:
- Pin 8 - 48.0
- Pin 7 - 20.1
- Pin 6 - 29.5
- Pin 5 - 29.5
- Pin 4 - 0
- Pin 3 - 24.9
- Pin 2 - 24.9
- Pin 1 - 38.9

IC2:
- Pin 8 - 40.9
- Pin 7 - 23.9
- Pin 6 - 33.5
- Pin 5 - 33.5
- Pin 4 - 0
- Pin 3 - 33.5
- Pin 2 - 33.5
- Pin 1 - 21.9

SSM2212:
- Pin 1 - 33.5
- PIn 2 - 29.6
- Pin 3 - 28.9
- Pin 4 (NC) - 40.2
- Pin 5 (NC) - 38.8
- Pin 6 - 28.9
- Pin 7 - 29.3
- Pin 8 - 33.5

CHANNEL TWO (reading ~25 V at TP3)

IC1:
- Pin 8 - 48.3
- Pin 7 - 30.9
- Pin 6 - 29.4
- Pin 5 - 29.4
- Pin 4 - 0
- Pin 3 - 23.8
- Pin 2 - 23.8
- Pin 1 - 27.9

IC2:
- Pin 8 - 41.0
- Pin 7 - 22.6
- Pin 6 - 33.6
- Pin 5 - 33.6
- Pin 4 - 0
- Pin 3 - 33.6
- Pin 2 - 33.6
- Pin 1 - 22.9

SSM2212:
- Pin 1 - 33.6
- PIn 2 - 29.5
- Pin 3 - 28.8
- Pin 4 (NC) - 39.5
- Pin 5 (NC) - 39.6
- Pin 6 - 28.8
- Pin 7 - 29.5
- Pin 8 - 33.6

The biggest abnormality I see is pins 1/7 of IC1 for channel 1 are clearly not making sense compared to channel 2. Everything else looks reasonably balanced from channel to the other to my eyes.
Have you tried the standard BC transistors in one BAN to see if it works fine?
No since I thought the SSM2212 was a standard substitution but with possibly higher performance, and it was passing sound find I assumed it should work with that. I am planning to try that for my 3rd channel build which I will start next.
Also email me some large well lit clear in focus photos of the top and bottom of the BAN PCBs.
Done.
There maybe something wrong with the supply or it is has to small a load on it. These SMPS get quieter with more load. But 2 or more channels should be fine.
Just received a replacement one for $20 off mouser since I was placing an order for parts anyway and sadly it's not making any difference. The fluctuating noise floor persists, which tells me maybe there's something wrong with using the SSM2212 in this build?
Please answer all questions and send me the photos so I can work out your problem.
I hope so. I wish I had just paid you to build them for me. I've never had problems like this with a build before.
Last edited by mike on Fri Nov 28, 2014 11:10 pm, edited 2 times in total.

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BAN Build Thread

Post by Joe Malone »

Can you please check your IC pinout as pin 4 should be at 0v on both IC's and pin 8 should 48v one IC and about 40v on the next? Otherwise you have the ICs in backwards
Joe :-)
JLM Audio
Capturing Audio without Injury

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

Joe Malone wrote:Can you please check your IC pinout as pin 4 should be at 0v on both IC's and pin 8 should 48v one IC and about 40v on the next? Otherwise you have the ICs in backwards
Sorry Joe, you're right. I rechecked and I have my ICs oriented correctly, but I reversed the pin numbers when I was writing down my voltages. I've corrected them.

Again, it appears like the TP3 problem with channel one is originating at IC1 pins 1/7 (output), as all voltages prior to that point appear roughly the same and +/- balanced between the two channels.

mike
Posts: 18
Joined: Thu Oct 23, 2014 11:14 am

Re: BAN Build Thread

Post by mike »

To summarize, my problems appear solved after corresponding with Joe via e-mail, and the explanations are as follows:

1) Fluctuating noise floor - This has resolved to normal limits after adding a third channel to the build. It appears this was a shortcoming of the 48V adaptor that causes it to produce increased power noise at low current draw. I expect this will only improve further once I add the 4th channel.

2) Abnormal voltage at TP3 - Joe has clarified that TP3 voltage only indicates the balance between the two opamps which is not significant. I calibrated by unit with the following procedure instead:

(i) Measure voltage at the pin of 10K trimmer towards the front of the chassis, adjust 10K trimmer until this test point reads 24V.
(ii) Adjust 1K trimmer until voltage across gain pot is as close to 0.0 mV as possible.

Results
My specs running RMAA through the device with the pad engaged and the gain pot all the way down have been interesting. I currently have two channels with SSM2212 transistors installed and 1 with stock BC560's.

Many people have switched to the SSM2212 from other transistors for 990 type opamps saying these offer the best performance for the price. However, at least based on the three channels I have this does not appear to be the case.

SSM2212 Channel at Zero Gain:
Freq Response: +0.03 -0.05 db
Noise level -107.1 db
Dynamic range 107.0 db
THD 0.0015%
IMD+noise 0.0027%
IMD+noise swept 0.0033%

BC560 Channel at Zero Gain
Freq Response: +0.03 -0.05 db
Noise level -106.4 db
Dynamic range 106.4 dB
THD 0.0011%
IMD+noise 0.0018%
IMD+noise (swept) 0.0022%

Thus at zero gain, the BC560 channel is producing far less IMD, and less THD as well. This was consistent compared to the other SSM2212 channel (which had 0.0018% THD). The BC560 channel on the other hand has ~0.6 dB higher noise than the SSM2212 channels.

SSM2212 Channel at Max Gain
Freq response +0.05 -0.06 dB
Noise level -69.2 dB
Dynamic range 69.2 dB
THD 0.0097%
IMD+noise 0.097%
IMD+noise (swept) 0.101%

BC560 Channel at Max Gain
Freq response +0.04 -0.05 dB
Noise level -68.9 dB
THD 0.0095%
IMD+noise 0.099%
IMD+noise (swept) 0.103%

When testing the channels at full gain, the results were almost identical however with no clear differences.

Remaining Questions
Joe I'm wondering if you can comment on a few things:
- Are these results normal or less than optimal?
- Does it "make sense" that the SSM2212 would have higher distortion but slightly lower noise?
- Do you think an LM394 would be superior to both? On a previous post you suggested it can provide 1-3 dB noise floor benefit over the BC560.

Thanks,
Mike

User avatar
Joe Malone
Site Admin
Posts: 2069
Joined: Wed Jan 24, 2007 11:35 pm
Location: Brisbane, Australia
Contact:

Re: BAN Build Thread

Post by Joe Malone »

Remaining Questions
Joe I'm wondering if you can comment on a few things:
- Are these results normal or less than optimal?
Yes in this test setup.
- Does it "make sense" that the SSM2212 would have higher distortion but slightly lower noise?
Not a big enough number of test sample to be sure but the higher current drawn by these multi transistor matched pairs mean they have emitter share resistors to help the matching which can cause this. But THD for all is so low it matters little. The main feature is usually best noise floor which multi matched pairs can give 1 to 3db theoretical improvement which is shown in the results.
- Do you think an LM394 would be superior to both? On a previous post you suggested it can provide 1-3 dB noise floor benefit over the BC560.
While SSM2212 and LM394 give a slightly better noise improvement I have always found I like the sound of the BC560 pair better to my ears.
Joe :-)
JLM Audio
Capturing Audio without Injury

Post Reply