904a Ladder Filter

[analogmonster]

Hi,

I'm just on cloning the 904a Low Pass Filter and I have a question. Can anybody tell me the resistor values of R25 (trim pot) and R45 (R.47 ???)of the 1967'er schematic? (See red circle below).

904a_unknownResistors.jpg

The resistors are located in the current sink below the first filter pole.
Perhaps one of the lucky owners of an original module can help me.

Thanks in advance

Carsten

[Kevin Lightner]

It's a 2.5 ohm wirewound trimmer across a 0.47 ohm resistor. (Yellow, Violet, Silver)
Since resistors like this are usually used with some real current, they may be difficult to find in small packages.
Most 904As used large ones. I'd guess 1 or 2 watts.
Also if it doesn't say in your schemos, the resonance (emphasis) pot must be a 50K reverse audio taper pot.

[cloudhop]

Hi Kev

In my 2 x 904a clones I built from the schematics , I have the values at : R25 = 2.5R and R45 = 0.47R.... not sure where I got the info from though as it was about 10 years ago..but I know the R11 piggy back proved a problem. I couldnt get the suggested values to work at all.

[Kevin Lightner]

Note sure what to suggest.
I've seen Moog schematics missing parts, parts values, have diodes or caps drawn backwards and other errors.
I don't suggest anyone try and build some of these old machines using the schematics alone.
It's best to compare them to the real thing, if available.
I have a couple of 904A filters here now.
Also various photos and scans from previous 904A filters.
Would pics help?

Are you using original parts numbers?
Sure of the transistor pinouts?

[analogmonster]

Note sure what to suggest.
I've seen Moog schematics missing parts, parts values, have diodes or caps drawn backwards and other errors.
I don't suggest anyone try and build some of these old machines using the schematics alone.

No. It doesn't work, as my experiences with my first 904B clone show. I completed missing details and replaced wrong information by using LTSpice and by drawing a spice model of the 904B. I tested my clone until it worked accurate before I soldered just one pin of the second prototype. Then I built it physically and had to change only a few details to get it into work in the real world.
Of course this way does not show me how close I am to the original, but as I use actual components and I don't know whether these work with the original circuit I had to go this way or I would have to build tons of prototypes instead.

It's best to compare them to the real thing, if available.
I have a couple of 904A filters here now.
Also various photos and scans from previous 904A filters.
Would pics help?

Yes, please. All information / photos / scans can help. Especially part positions and corresponding layout pics would be fantastic.

Are you using original parts numbers?
Sure of the transistor pinouts?

No, as mentioned above. But I know the old pinouts of course (data sheets).

Thanks in advance

Carsten

[analogmonster]

I made spice simulations of the circuit with and without R25 / R45 (meaning the emitters of Q7 and Q8 are directly connected to R23). According to spice there is no difference in signal behaviour.

Carsten

Hi,

I'm just on cloning the 904a Low Pass Filter and I have a question. Can anybody tell me the resistor values of R25 (trim pot) and R45 (R.47 ???)of the 1967'er schematic? (See red circle below).

904a_unknownResistors.jpg

The resistors are located in the current sink below the first filter pole.
Perhaps one of the lucky owners of an original module can help me.

Thanks in advance

Carsten

[Kevin Lightner]

Not sure what to say.
SPICE simulations can often be misleading or plain wrong.

For example, a circuit might "work" fine without a part somewhere, but then the current draw is far higher once built or parts blow out.
Sometimes loading, DC offset or dynamic impedance changes are not allowed for, considered or checked.
I have all sorts of circuits that won't work on Electronics Workbench, but work fine in real life and vice-versa.
(Try making a reverse avalanche noise source in SPICE)

Anyway, I still have a couple 904As here, but they'll be leaving soon.
But I'm currently not financially able to take all the time necessary to scan or take quality photos of both sides of each board without some sort of compensation.
904As require some disassembly to gain access to the trace sides of their two boards.
Perhaps someone else can step up and help for free?

[MC]

Have to agree with KL on the Spice inaccuracies. Good for education but in the real world the component models leave a lot to be desired.

[analogmonster]

Not sure what to say.
SPICE simulations can often be misleading or plain wrong.

For example, a circuit might "work" fine without a part somewhere, but then the current draw is far higher once built or parts blow out.
Sometimes loading, DC offset or dynamic impedance changes are not allowed for, considered or checked.

Hm. You are right. I blew up a multiturn trim in my other filter module, the 904B because I did not check the current at one of my modifications of the original circuit. And I had this DC offset thing which I had to compensate. So I checked the current through R23 of the 904A now and spice says 4 mA, so I hope it will be ok for the real prototype.

Anyway, I still have a couple 904As here, but they'll be leaving soon.
But I'm currently not financially able to take all the time necessary to scan or take quality photos of both sides of each board without some sort of compensation.
904As require some disassembly to gain access to the trace sides of their two boards.
Perhaps someone else can step up and help for free?

No Problem. I build the 904B prototype now according to my spice model, and we will see how close it works to the model.

Carsten

[EricK]

You would be wise to listen to these guys.

[Kevin Lightner]

You would be wise to listen to these guys.

Thank you number one poster for adding your informed opinion.

I blew up a multiturn trim in my other filter module, the 904B because I did not check the current at one of my modifications of the original circuit.

Happens.
I tend to power up prototypes and modules on a programmable power supply.
Has digital readouts displaying current draw live, voltages, over current, etc. Keypad entry.
You can limit the current from 1 ma to 3 amps, so I tend to dial in modules in the normal current ranges expected.
Obviously a 960 sequencer takes a lot more current than an envelope generator, so I provide only what something is expected to draw.
If there's a short somewhere, there may not be enough current (hopefully) to do any real damage.

My power supply model is no longer made, but there's many like it with programmable dual and triple outputs.
This one actually has a communications port too.
I could conceivably send it pre-programmed setups for various makes and models of modules, but I've never bothered.
I have plugs (jigs) that plug right into it for Moog, Emu, Sys-700, Polyfusion, Oberheim SEMs and others, so it's really easy to punch in the correct values and plug whatever module right in for testing.

I have a sheet somewhere here of approximate current draws for most Moog modules.
But they can vary with rev and I know my list is incomplete.
The 902 VCAs had at least 5 different revs made. All different PCBs and circuits.
Seems Bob was obsessed with making a good VCA.

Fwiw, a IIIP here now is only drawing about 950ma on 12 volts and approx 275ma on -6 volts.
That can vary a bit as envs are gated and outputs loaded, but even a whole IIIP basically takes an amp, lamps excluded.
You could power one off solar cells nowadays.

[analogmonster]

Of course I meant building my 904A prototype. My 904B is already up and running

No Problem. I build the 904B prototype now according to my spice model, and we will see how close it works to the model.

Carsten

I must admit that I don't have too much experience with spice (LTSpice to be more precise). The first real success was building the 904B High Pass Filter according to my spice model of the original circuit. The CV mixer section in my physical prototype worked exactly as analyzed in the spice model; some dc offsets occured in the filter section, but the signal behaviour was identical again compared to the model. So I hope the 904A prototype will work as well. We'll see...

[EricK]

Thank you number one poster for adding your informed opinion.

You're welcome, Friend.

Eric