Satyrnine
-
Posts
127 -
Joined
-
Last visited
Content Type
Profiles
Forums
Events
Posts posted by Satyrnine
-
-
I'm going to put together a mouser BOM for the ubaltobaltube shortly, unless someone has one they could share?
-
20 hours ago, justin said:
Been working on this for a looong time. CFA3 from scratch mixed smd/thru-hole design. Currently assembled with NOS Toshiba 2SC3381/2SA1349 inputs, NOS MPSW TO-92L thru-hole transistors and bcxxx for the smd. Power consumption 75W no fan. Linear PSU with 20,000uf/63V filter capacitance — custom golden reference regulators. Final dimensions are 12x9x4.5” / 22lb+
the motorized 48 step attenuator w/ remote and display was a last minute addition. definitely interested in finding out the feedback for that, but would want it built backwards for a final design, I think.
A friend is at canjam right now and just sent me a photo of this beast there. Great work!
-
1
-
-
Well, the problem is I’m an idiot apparently. After going over the board carefully so many times, I came back to it after a breather and there it was, Q8 in backwards. I’m so sorry for wasting everyones time with this. 😫
-
1 hour ago, simmconn said:
Is it the 'ok bias' channel in your schematic that has large output offset with servo disabled? I gave my analysis in my post. Q24 and Q25 form a current mirror. The current through R21 should be the same as the current thru R31 and thru the R26. Apparently it's not the case based on your measurement. I would focus on the related few components.
Ack, I apologize simmconn, yes its the “ok” board. I’ve been wracking my brain on this and failed to recall you’d already pointed out that the OK is def not OK. Thank you for your help, will investigate there, makes perfect sense.
-
1
-
-
1 hour ago, dip16amp said:
What is the offset without the op amp? It should be around 40mV. Then the op amp will bring it down to near zero.
Don't think the transistors hfe would be the cause of high offset.
What are the other boards offset without the op amps?
Without opamp, about 400ma 😬 The others are around 40ma as expected.
I Just replaced all the output trannies, the input pair, and the ptza’s with as close matched parts as I could, no luck.
-
Trying to diagnose some excessive offset on one board. Likely culprit is unmatched N/P pairs I assume. I ordered 20x spare of each bc556 and bc546 but there's no way I can match these. 546's are all about 215hfe, and 556's are all about 120hfe. Same with ptza06 and ptza56, at least 50hfe difference between N and P types, but tight matching within their own type.
I replaced the input pair and the ptza's with the closest hfe match I could find, helped offset ever so slightly, maybe 20ma.
Should I be matching vbe instead of hfe? Sorry, tube guy learning the ropes.
-
2 hours ago, mwl168 said:
Assuming all the removed transistors are indeed all good, this leads me to wonder if the original issue was caused by cold-soldering?
It's possible, although my soldering skills are pretty developed. I like use Kester 186 liquid flux, Kester 44 60/40 solder and a decent Hakko Iron at 750*. Taught by a guy who helped make some electronic parts for the F117 fighter as a civilian contractor even. Everything is always nice and shiny and liquid looking. I wonder if like tubes, a tester only tells you a part of the picture. I've had numerous tubes test OK on a tester, but not in circuit.
-
1
-
-
Swapped all 15030/31’s and the bias issue is solved! Behaving nicely. I tested all the removed transistors with a cheap ebay tester and none tested “bad” but I realize thats a limited test. Thanks for the help all. I can finally try this beast in balanced mode!
Still waiting on tube input boards from Dukei who had spares. Also need to wrap up dc connector details and order chassis yet.
-
2
-
-
1 hour ago, simmconn said:
The protector board needs a solid ground, i.e. split +/-15V power with their common GND tied to the amp GND. A single floating 30V would not work, because the 12V relay only draws power from the +12V and will throw the GND off center and causes feedback to the comparators. You can share the +/-30v with the amp. Make sure the variety of 7812/7912 you use are rated for 35V or more, and the 7812 has sufficient heat sink for continued operation of the relay.
I'm using a +/-30v grlv with common ground to power it. Tried it shared with a CFA2 as well as on it's own GRLV. I still get the noise both ways. Only way I don't get the noise is disconnecting cfa2 boards, powering only the protector board and headphones connected to it's outputs.
I'm using onsemi 7812 and 7912, rated for 35v input and 1A. Relay is Omron G6A-234P-ST-US-DC12 which is 12v and ~17ma coil current.
I'll check on heat, but seems if it was the issue, I'd get noise even without CFA boards connected, no?
Thanks!
-
3 hours ago, simmconn said:
@SatyrnineWhat does it mean by the minus sign on the voltages? It wouldn’t make sense unless you mark which pin had the red probe on it and which one had the black one. You might instead mark the actual polarities, like the pin with higher voltage with a (+). For voltage to ground such as the servo output, it should be the actual voltage. You marked -10V on the schematic but said 10V in the post, which one is true? Based on the voltage drops on R19/R21, it should be 10V, so the lower current mirror should have more current than the upper one. But that’s not the case when you compare the voltage drops across R25 and R31. So something must be off there. The transistors seem very mismatched. Btw, most of the reference designators are hardly legible on your hand-marked schematic. If you expect people to help you, at least make it easy for them to do so.
You're completely right, arbitrary to have a negative since I was just measuring drop across the resistors. Dukei recommended measuring drop across resistors vs voltages to ground. I mindlessly added the negatives when I was jotting it down, just a matter of probe position in this case. +10v is the proper voltage after the servo in this case.
Thanks for the input, I'll investigate further. Sorry about the size. The resolution is decent, but on a phone or such device it may be tough. The schems I was jotting on were small, tried to keep everything visible. I def often write too small.
-
1 hour ago, Pars said:
I take it that you have checked to make sure all components are the correct value and in the right place? From your voltage measurements, everything looks more or less identical until you get to the servo area. What are the voltages on the inputs to the opamp?
Also, in your picture previously, 2 of your output devices appear to be missing screws into the heatsink.
At this point, I would probably replace all the 15030s/31s given your probe slip just to make sure. These boards are pretty easy to rework.
Yes, I've visually verified all parts, and ohmed out each resistor to verify. Checked for cold solder joints, polarity/orientation/etc. N's and P's in right spots.
Good eye Pars! Yes, in that pic the two toasted output devices had the screws removed, I've since replaced with new official distributor parts. I do have spares, I'll just replace all the 15030/31's as you suggest. Those are easy as you say, esp with my new Hakko desoldering gun.
Inputs to opamp would be the feed/sense from the output right?
I should add, I have the amp working in SE mode now, with 120ma bias or so, and it absolutely blows my mind. Unreal power and slam, and the dynamics and detail are just wild. Raw dogging for now without protector board with my LCD-2's and LCD-X's.
-
Still haven't been able to test at below 30v, BUT I did find that if I have no CFA boards connected, just the protector board powered, I don't get any hum from headphones. I also built up a second board, and had same symptoms as first, so I'm guessing it doesn't like 30v.
-
I bit the bullet and took measurements of an OK and the "Hot Bias" CFA2 boards. Here is what I found:
Test conditions:
OK board biased to around 120ma, trimpot in middle-ish position.
Hot bias board has trimpot all the way down, which gives about 100ma of bias.
Ran from the same GRLV at 30v.Clearly the big difference is the output of the servo is showing 10v on OK board, and only .8v on hot bias board. Servos getting proper voltage on each board. I swapped in a known good chip, no difference. The 5k feeders after servo are showing dif drops between boards too.
Can anyone give me some guidance on where to go from here?
-
I ended up using a spare 30v grlv to power the protector board separately. Same transformer. +/- 12v on pins 3/12, still had noise, but it was different sounding this time maybe. Less hf content? Maybe imagination. That mostly (?) rules out that it's somehow injecting the noise back through the grlv and through the cfas, but solidifies that it's happening at the protector board. I think. Suspecting overworking the 12v regs w 30v at this point.
Wasn't able to test at lower input voltage because my bench supply isn't bipolar apparently. (I usually do tube stuff, rarely need a bipolar supply until now, hah!)
I'm going to go through it and triple check values and get some more detailed readings.
-
I do have a bench supply, will def give that a shot and report back. I'll try it at 12v and 30v to see if there's any difference in behavior.
-
Will have to check the pins tmrw, but fwiw, spec sheet says the regs used max input voltage is 35v, so it is cutting it close. And yes, all unused input pins grounded. On delay is only a few seconds too. I'll triple check everything again asap.
-
I built a protector board up for this. When I connect it to one of the grlv set to 30v (also powering 2x CFA boards), it causes pretty bad high frequency noise on the outputs. Right now in SE mode, grounding negative inputs.
I tried grounding two ways:
1) GRLV ground AND cfa output grounds all connected to protector board. (hf noise was present)
2) Only GRLV ground connected to protector, cfa output grounds and protector input ground connected to star ground. (hf noise was worse)
When I remove the protector, cfa is dead quiet. I've read these can be used with 30v, but maybe I'm stressing 12v regs too much?
-
I built a protector board up for my 30v CFA. When I connect it to a grlv set to 30v (also powering 2x CFA boards), it causes pretty bad high frequency noise. Right now in SE mode, grounding negative inputs. Negative outputs not connected to anything.
I tried grounding two ways:
1) GRLV ground AND cfa output grounds all connected to protector board. (hf noise was present)
2) Only GRLV ground connected to protector, cfa output grounds and protector input ground connected to star ground. (hf noise was worse)
When I remove the protector, cfa is dead quiet.
Thoughts?
-
On 5/16/2023 at 4:39 PM, audiostar said:
The Amphenol's don't have removable pins, so no crimping possible and also no variable size pins available as well. It only has one 18 AWG pin (and all the others are max 20 AWG wire pins), so not really suited for a high current class A amp.
A much better connector is the AMP CPC shell Size 17, p/n 206043-1, 206044-1 (PSU side), 183040-1, 183039-1 (amp side), as used on the BHSE or the Dark Star. Pins are removable and all sorts of pins for crimping (as well as soldering if you wished) and sizes for 16...26 AWG wire are available for the same size 17 connector. Very nice modular and serviceable system. It is HV rated and the 16-18 AWG pins would be perfect for the beefy GRLVs. Really nice.
Was going to go with switchcraft due to being less modular and worried about getting the right pins and all, but these are WAY cheaper than switchcraft too. Do you happen to know the right "pins" I'll need? I'd hate to order the wrong thing, and there are TONS of options with this line. Are locking nuts, plug housing, and everything but pins included with the connectors, or do those need to be sourced separately too? Doesn't seem clear. Datasheets show no accessories. Thanks!
Also, does anyone have any guesses on what could be causing the spicy bias on the one board? I'd really hate to resort to shotgunning all the sand to try to find the culprit. I have a decent Hakko desoldering gun, but extensive rework sucks if avoidable. I'm going to go through and measure voltages across each resistor and compare to a working board.
-
29 minutes ago, jamesmking said:
for smd and transistor probing I highly recommend spring loaded very fine tip probes. I use these
https://probemaster.com/spring-loaded-micro-tip-test-lead-kit/
the tips are replicable and have an assortment of end types and the leads are high quality...
for more hands free measurements I use Sensepeek 4011 PCBite 2x SP10 Probe set
https://telonic.co.uk/product/4012-sensepeek-pcbite-kit-with-2-x-sp10-probes-for-dmm/
they are not high voltage rated because of the exposed metal and relatively thin insulation on the wires but they stay in place. The tips are as sharp and fine as the probe master. They also sell scope probes:
Oh! Excellent, love both options! My main work is with instrument tube amps which often have a bit more real estate, so my probes are not so small. I think part of my issue is there was no good place to rest my arm/hand to stabilize as I probed, since this is all out on the table in spaghetti form currently.
-
Any reason I can't use this with a 30v balanced CFA2 if I plan on driving small speakers with it? It's rated for 1A, 2A max, and 30w at 8ohm is right at 2A. Thoughts?
-
Last CFA board has spicy bias. With bias pot all the way down, I get about 150ma across the 1R's. That can't be right. The rest needed to be set around middle to get any idle current at all. Rechecked my P's and N's, and resistor values. Any thoughts? Suspect the middle (bias driver?) 15030? 🤔
Edit: replaced the middle 15030 with a slightly lower hfe version, dropped minimum bias to 120ma or so, so not the culprit. And then I shorted legs with my probe again like an idiot and popped the "last" 30/31's again. You'd think I would have learned something. I need probes designed for this clearly.
Here's latest chassis face design:
The horizontal lines would be deep machined grooves.
-
On 5/22/2023 at 11:19 AM, dip16amp said:
Just replace both 1 ohm resistors and the one bad 15030. Everything else should be fine.
Measure resistance across the 15031 and 15030 to find a shorted one.
I've only had a bad 15031.
Measure resistance across both resistors may be 1.5 ohms instead of 2 ohms.
Measure these in circuit with power off.
Thank you! This is great to hear. Was concerned it may have taken out more, I will measure everything! The one blown 1R is currently about 250ohms, haha! Will replace both though, as they both got to warm.
-
5 hours ago, Satyrnine said:
Fired up first two cfa2's last night. A couple issues:
Board 1) Without servo, I'm seeing around 400mv (!) offset on output with input shorted. If I plug in the opamp, it slowly comes down to 1-2mv over 30sec or so.
I started with bias pot all the way counter clockwise. Not seeing anything across the 1R resistors after a good 8 turns clockwise, do I just need to keep going, or should I be seeing some idle current by that point?
Board 2) LED's not lighting. It was late, and I gave up before investigating further.
Board 1) Butterfingered probing one of the 15030's trying to diagnose the high negative DC offset on this board. Shorted base to connector momentarily, 1R's went nuclear and started melting my probe tips within 1 sec. I powered down a second later, but I'm guessing I have some sand and resistors to replace now. Any suggestions on what to replace? Neat. Stupid mistake.
Board 2) Remind me not to trust mouser to always get their parts picking correct. Could have been my mess up too I suppose. There was a 10k in place of a 10R near the input (R9 or R10) which I deduce was causing the LED's not to light. Replaced that resistor w correct value, LED's light and this board is now showing around 40mv offset without opamp and zero offset with opamp, which seems about right. Biased to low around 80ma for now, just had to adjust trimpot further. This board seems to be good to go.
3 hours ago, audiostar said:400mV is around 10x too much for the offset.
8 rotations are normal for the bias pot until you start noticing the voltage increasing. The pot has around 30 turns or so. 150 to 200mV would be in range.
Indeed too much. Now that I messed this board up, may be replacing a bunch of sand anyway. I double checked my N's and P's and resistor values, and polarities before it went south, everything seemed to be in the right place.
and now for something completely different part 3
in Do It Yourself
Posted
Does anyone have any reading/data/thoughts of case temperature of the 15030/1's at various bias levels? I'm on temp less than ideal heatsinks, but direct mount with thermal paste. Getting about 50*C across the board after and hour or so with 30v supply and around 120ma bias. (Planning on 200ma once I'm on the dissipante sinks) Using a 4 channel thermocouple.