MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Croglin filter doesn’t filter
MUFF WIGGLER Forum Index -> Oakley Sound Systems Goto page 1, 2  Next [all]
Author Croglin filter doesn’t filter
ProducerMatt
Hey folks. I bought and assembled the Croglin kit from Synthcube (I didn’t know how to find parts at the time I ordered it) and I’ve run into a snag: it doesn’t filter. The trimpots seem to make no difference. The only pots that change anything are the resonance (adding a high end buzzing, which makes me think the filter is there but open) and duality, which adds and subtracts as (I imagine) it should — I notice that it doesn’t cancel itself out completely, maybe that’s not normal?

I’ve thoroughly checked for shorts & backwards parts. I haven’t yet checked for incorrectly placed parts, as the Croglin had all parts in separate bags with the corresponding #s labeled, so I feel that’s not likely. (Also, I can’t sightread resistors, so that combined with the ordering of the Oakley partsheets makes the process misery.)

For inter-board connections, the Synthcube kit came with those TE Connectors that are IDC. The result looks pretty gnarly (MASH MASH MASH), but I’ve checked all the pins (on the bottom of each board) and they’re connected just fine.

Despite my embarrassment, I‘m going to upload some photos of the board so your significantly more experienced eyes can take a look. In the meantime, let me tell you my test environment.

I’m currently using a (relatively-cheap) power supply set up in series mode. My signal ground is the + of the negative supply and the - of the positive supply. The panel ground is the green earth ground. I also tried using the earth ground for both grounds, which didn’t work.

This is the first module I’ve built, so I’m using a Little Phatty as an input signal, and I’m sending the output to some cheap PC speakers.
Synthbuilder
ProducerMatt wrote:
I’m currently using a (relatively-cheap) power supply set up in series mode. My signal ground is the + of the negative supply and the - of the positive supply. The panel ground is the green earth ground. I also tried using the earth ground for both grounds, which didn’t work.


This grounding has me confused. Forget mains earth for the moment.

If you are using the four pin MOTM style connector, PSU, then both the module's 0V pins, that is pins 2 and 3 of the PSU header, need to be connected to + of the negative supply and the - of the positive supply.

If you are using the five pin dotcom connector, PWR, then pins 2 and 3 should be linked on the PSU header.

This will ensure that the ground lugs of the input and output sockets are connected to the same 0V as the module. Without this the module is unlikely to work.

Check the voltages on the power supply with respect to 0V. That is make sure the voltage reaching the board is indeed +15V and -15V with respect to 0V at pin 2 (or 3) of the PSU header. +15V should be found at pin 8 (top right pin) of any of the TL072 op-amps TL072. -15V should be found at pin 4 (bottom left pin) of any of the op-amps.

It's worth trying to learn the resistor code. It's actually quite easy although the hardest part is working out which way round the part should be to read it in the correct order. To test your reading accuracy check the resistance with a multimeter. Although, once the part has been fitted into the board it's not possible to measure the resistance reliably.

Tony
ProducerMatt
EDIT: Oops! Didn’t see your post, Tony. Sorry. I’ll try everything you suggested.

Here’s the photos. https://imgur.com/a/rCPic

I know my environment is messy, but it’s a lot better then it was. I clean it up a little more every week, maybe one day it won’t make professionals throw up on sight! Mr. Green

I did a check on my power supply and found there was 60hz leaking through by almost a volt P-P on all terminals. From the day I first got that supply, I knew it would stab me in the back. Flamey
ProducerMatt
Hey Tony. Just tried your suggestions, and found that +-15V was at all correct spots on the chips. The board doesn’t behave any differently than it did with signal ground connected to earth.

Tomorrow I’ll try hooking up my 15V switching PSU I bought for the modular to see if I still experience these issues. If that’s a bad idea somehow let me know.

Regarding theory: I thought that the jacks needed to be connected to earth so that the different 0v levels wouldn’t all screw with each other?
JohnLRice
Hi ProducerMatt,

For what it's worth in November I built a Croglin from a SynthCube kit and it works well. Here are some pictures of it. (note that I was missing the correct value for R51 so I substituted two resistors in series that i had in my parts bin)






Synthbuilder
ProducerMatt wrote:
Regarding theory: I thought that the jacks needed to be connected to earth so that the different 0v levels wouldn’t all screw with each other?


The metal chassis of your modular, and the sockets that connect directly to it should be connected to earth. This is ideally done directly and if you are fitting a mains power supply into your modular case there must be a direct connection from the mains earth to the external metalwork (including the front panels) of your modular. So once you fit that switch mode power supply you linked to in your post you will need a mains earth to be fitted to both the power supply's earth point, the case metalwork and the module's 0V.

To reiterate: Do not omit a direct earth bond if you have a mains powered power supply inside your modular. Power your case with a three core cable that supplies both the mains power and an earth.

Now, it is also possible to allow your modular to float with respect to earth if you are using an external power supply, like a wall wart mains adapter, that is labelled as double insulated. A double insulated device means that it has been designed that, in all expected cases of failure, it will not allow a dangerous voltage through to its outputs. Using this sort of supply means there no dedicated connection from mains earth to your modular. The metalwork will then only be indirectly earthed by any earthed equipment that is then connected to it. This could be your synth, mixing desk, audio interface or what have you. The problem is that this is not always going to happen. For example, your audio interface is maybe connected to a laptop running from batteries and you are listening to it on headphones. This isn't dangerous so long as you are powering your modular with a double insulated power pack. And it will work fine in most cases. I normally recommend doing it this way (see Oakley PSU) simply because using a sealed external power supply keeps the dangerous high voltages out of the modular case.

The problem comes when you have a large music set up with multiple sound generating boxes and you are trying to connect them all together. This is when you want a low resistance earth connection between all your devices. You can't always rely on the wimpy audio cable to do this for you. However, in many cases it does work well enough for people to produce great music.

Generally the simpler your set up the fewer problems you will have. Add lots of unbalanced audio or CV connections and the hum and interference problems start to mount up.

Tony
Synthbuilder
I notice from both sets of pictures that Synthcube is supplying 5% and 1% resistors. This does make it confusing since both types of resistors have different colour codes - well, it's the same code just different numbers of coloured rings with the 1% ones having five rings not four. Personally, I always recommend sticking to just 1% metal film types and learning one set of codes.

Tony
Synthbuilder
Faultfinding:

Check the voltage at pin 7 of U1. Wiggle the frequency pot. The voltage should vary between around +7V to -7V. If not suspect a problem around U1.

Does pin 1 go to pin 1 on the 8-way interconnect that goes to the socket board? It's difficult to see on the photographs. When you take them out of the sockets and hold both sockets downwards the wires going between them should not cross over but travel straight like tram lines.
ppkstat
My recommendation is not to use a switching PSU. Analog circuitry can be quite picky with it's power and with the particular one you're thinking you will get none of the advantages (power, 1.3 A is not a lot for a switching PSU) but you will get the main disadvantage (noise). The only thing that will be better with this psu in comparison with a linear one will be reduced heat generation due to higher efficiency.

The Oakley PSU is a very good design and there are others as well.
Synthbuilder
ppkstat wrote:
My recommendation is not to use a switching PSU.


I agree. Although, the Sanken HWB one does seem quite good for its type. 5mV noise is probably OK although it doesn't state what level the 'spike' is. Indeed, the datasheet seems remarkably lean on data.

There's a better one here:

http://www.sanken-ele.co.jp/en/prod/library/pdf/k1-s11ea0.pdf

But again somewhat lean on real performance data.

It'll not be as quiet as a linear PSU but it may be good enough. Correct placement within the modular case and proper earthing will be required.

Tony
ProducerMatt
Tony: the Freq pot gives Pin 7 of U1 a range from +12.31 to 0. So clearly that’s not how it’s supposed to be. very frustrating Guess I am gonna check these resistors after all.

I’m really confused about this connector. The wires do cross, but on my setup I don’t think they could NOT cross?

On the main board, the side of the connector with “I/O” written on it is the left, pin 1, where the Audio In runs, yes?
and opposite is pin 8, ground, yes?
And on the socket board, the side near L1 is pin 8 ground (on the left looking down on it), and the opposite is pin 1 Audio In, yes?

The traces sure look like that’s the case. And if so then I can’t not cross them. I went over this so many damn times. Also, I’m getting audio so I don’t see how I could have them mis-wired.
Synthbuilder
ProducerMatt wrote:
Tony: the Freq pot gives Pin 7 of U1 a range from +12.31 to 0. So clearly that’s not how it’s supposed to be. very frustrating Guess I am gonna check these resistors after all.

Hang on - my numbers aren't correct. Sorry. It depends on what the TUNE and SCALE trimmers are doing. You can indeed get the voltages you are seeing if your TUNE trimmer is set at one of its ends.

So that bit may well be working just fine. Just to check though; adjust TUNE all the way to its other end. This time pin 7 should go from around +7.4V to -5V.

Quote:
I’m really confused about this connector. The wires do cross, but on my setup I don’t think they could NOT cross?

Remove the whole interconnect from the module. Hold the interconnect so that in each hand you have one of the end connectors. Hold it so that both connectors face downward. You should have all the wires moving from one connector to the other in parallel and not crossing.

Quote:
Also, I’m getting audio so I don’t see how I could have them mis-wired.

Audio signals can sometimes travel along miswired connections and still come out the other end. This can lead to very confusing results. But I'm guessing here. Without the module in front me, hooked up and with a scope handy, all I can do is suggest things to check that may be wrong.

Just check the four transistors are in the correct places. Getting the BC560 and BC550 swapped will perhaps cause some frequency control issues. Q2 and Q4 should both be BC560 (or BC559).

What is the voltage on the top end of R36? It should vary when you change the frequency pot? What about the top end of R37? Does that vary similarly?

Tony
ProducerMatt
When I twist SCALE and TUNE all the way counterclockwise, I get 14V at 0 and a little less than 1V at 10. With SCALE clockwise, I get 11V to .7V. With TUNE and SCALE both CW, I got 6.6V at 0 to -4V at 1.

Quote:
You should have all the wires moving from one connector to the other in parallel and not crossing.


On mine, they are crossing. But, the ground connector is on the left on the connector board and the right on the main board, so they should be wired correctly unless the entire board is flipped? How is yours different? I don’t how it couldn’t be correct. This layout matches your photo and JohnLRice’s one as well.

ProducerMatt
I get -13.57V consistantly from R36 and R37, irrelevant of the freq pot. Maybe the transistors are the issue after all. I’ll check their model.
Synthbuilder
ProducerMatt wrote:
On mine, they are crossing. But, the ground connector is on the left on the connector board and the right on the main board, so they should be wired correctly unless the entire board is flipped? How is yours different?

It is a little difficult to tell on the photograph but I think your cabling is correct. When I say the individual wires should not be crossing this is only what the interconnect is unplugged from the module and stretched out between your two hands.

Tony
ProducerMatt
I’ve followed each wire individually and they go to the places they should.

I just checked the transistors, and the models are correct. Could they have been blown when I put the power in reverse?

(Still in the process of checking resistors, by the way. Haven’t seen any incorrect ones... YET spinning )
Synthbuilder
ProducerMatt wrote:
I get -13.57V consistantly from R36 and R37, irrelevant of the freq pot.


This is definitely from the top pads of each resistor? When I say top pads I mean the ones nearest the four transistors.

You should have these moving with the cut-off frequency. With the OFFSET pot set centrally, both should be the same voltage and both should move together. The higher the cut-off frequency the closer to 0V they both will move. At high frequency I would expect them both to be 0V. At the lowest frequency they should be around -14V.

What's weird is that they are not moving but also that when they are stuck at -13.57V, one should expect the LM13700 to be turned off and the Croglin passing very little audio.

I wonder whether the LM13700 is dead.

Tony
ProducerMatt
I’ll check the top pads this afternoon.

Do you have any suggestions for how I should check the LM13700? The only thing I can think is to build the datasheet’s example amplifier circuit 4 times, which would suck. Can I just probe the ins and outs while it’s getting a signal and see what it shows?
ProducerMatt
I just checked the top pads, and regardless of the frequency knob, they're both roughly around 0 volts even though the bottoms are -13V. I don't get it.

I'm not convinced I need to replace the transistors yet, but for future reference: What would you recommend for a transistor socket? Should I just get some of those pin sockets that are used on Arduinos?
Synthbuilder
ProducerMatt wrote:
I just checked the top pads, and regardless of the frequency knob, they're both roughly around 0 volts even though the bottoms are -13V. I don't get it.

Actually, that may be a good thing. I would expect 0V to open the filter which is what you appear to have. This suggests then the LM13700 is working. Pins 1 and 16 on the LM13700 are usually very close to voltage found on the negative rail.

That does suggest the transistors are not doing their job properly. The job of the transistors is to pull current down out of the LM13700 and the more current they pull the more open the filter. For some reason they are pulling the current down all the time. It'll definitely be worth checking that the transistors are correctly fitted.

Quote:
What would you recommend for a transistor socket?

I don't. Just solder them in and you'll have a much more reliable build. Transistors are not designed to go into sockets permanently. If you really do have to use them, then get turned pin SIL (single in line) sockets and cut them to size as they normally come in longer strips.

Tony
ProducerMatt
How do I know if they’re correctly fitted? Do you mean checking if the solder joint is good?
Synthbuilder
ProducerMatt wrote:
How do I know if they’re correctly fitted? Do you mean checking if the solder joint is good?

In the first instance check that the BC550 and BC560 devices have not been swapped. Q2 and Q4 should both be BC560 (or BC559). Q1 and Q3 should both be BC550 (or BC549).

If all seems well there is a good possibility that Q1 and Q3, or even all four devices, have been damaged by the power supply reversal.

Tony
ProducerMatt
It’s totally possible that they’re fried, but keep in mind that the filter behaved exactly like this *before* I reversed the power accidentally. So idk.

Still checking resistors. Mr. Green
ProducerMatt
Just realized I should let you know that I have a little 8-bit oscilloscope w/ FFT if you think that would be useful. Just in case you were assuming I didn’t have that tool available.
ProducerMatt
Okay, I just took one of the transistors out and checked it and it’s a dead short. I wonder if I fried it, or if it came broken. (Or maybe the thermal grease I put on it was electroconductive)

I ordered a component tester a while ago and I’ll probably be running every new component I get through it in paranoia.

Btw, I’ve checked, and all components are correct, including resistors.
MUFF WIGGLER Forum Index -> Oakley Sound Systems Goto page 1, 2  Next [all]
Page 1 of 2
Powered by phpBB © phpBB Group