CBS Model 450 Dynamic Presence Equaliser

[NOON]

I recently picked up one of these units quite cheaply. There's very little information out there about them and they didn't seem to be very popular back in the day, but they're very well built for reliable broadcast use. so I figured it was worthwhile even if it's just a parts donor.

I've traced out the circuit to draw a schematic because I can't find hide nor hair of one online and it's quite different to the related 400 series broadcast units. (Can't guarantee 100% accuracy, but I think it's pretty close.)

I can understand the input, gain cell and output stages pretty clearly, the CONTROL and DISCRIMINATOR circuits are a bit trickier, would appreciate some help in deciphering them.

I'll put up a series of posts with photos and schematic for discussion.

 

[NOON]

Schematic doc 1 - audio and control circuits. It's 5 removable modules in a chassis.

*UPDATED 26/02/2020 REV1.3 - Switch section redrawn, started labelling sections, sundry neatness corrections.
*UPDATED 8/3/2020 REV1.4 - Operational notes added. Final version, unless someone has any other corrections or notes to add.

 

[NOON]

Schematic doc 2 - Power supply. Pretty simple zener regulated PSU.

I find it a bit scary that the mains voltage comes in via the card edge connectors and the pins are only a couple of mm away from the low voltage DC lines and the fuse is on the card, so if it shorted to chassis or another line the fuse likely wouldn't blow.. Considering ways to make this unit safer.

 

[NOON]

The cards are populated with an unusual electrolytic capacitor type I haven't seen before. Not sure if they're aluminium or tantalum.
They're a white plastic encapsulated module and all still appear in great physical condition. I've only pulled one 100uF so far and it tested as 160uF and 0.6ohm ESR, so still seems good. Impressive for a late 60s unit.

 

[emrr]

Good news on those electrolytics.  The similar packaged black types found on a lot of mid-60s Gates are pretty much always bad. 

I had a later 1RU version for a minute and didn't find it did a whole lot. 

 

[NOON]

I don't expect it to be very exciting or worthy of rack space as-is, but considering it has some nice input and output transformers and a VACTROL gain cell and sidechain, I figured it wouldn't be too difficult to convert it into an opto compressor. Once I understand how it's doing what it's doing of course.

 

[mjrippe]

Glenn, those drawings are fantastic.  Are you an EE? 

 

[emrr]

Glenn, those drawings are fantastic.  Are you an EE?

No kidding, well done. 

 

[NOON]

Thanks. I studied about half of an EE course 25 years ago before being lured into IT for the money, but mostly it's because I also have training in draughting for building design. I thought they were pretty lazy sketches actually.

 

[PRR]

Thanks.

Maybe it's just late, but my mind wants to see the VGA as a subtraction bridge, signals crossed somewhere along the line:

Or maybe I misunderstand what this does.....

 

[NOON]

It's supposed to 'decide' whether there's enough presence around 4k-ish and if there isn't it will boost it. So kind of an anti-de-esser.

Entirely possible I may have (not) crossed the wires over when sketching things out if that's how you think it should go, I'll have a closer look at it later.

 

[NOON]

As I understand it, the opamp is summing the two 'phases'. The Vactrol bypasses extra signal around the resistors and the cap and inductor shunt together the higher and lower frequencies, cancelling them out and leaving just the presence frequencies to be added to the main signal. Please correct me if I'm misreading it.

 

[abbey road d enfer]

As I understand it, the opamp is summing the two 'phases'. The Vactrol bypasses extra signal around the resistors and the cap and inductor shunt together the higher and lower frequencies, cancelling them out and leaving just the presence frequencies to be added to the main signal. Please correct me if I'm misreading it.

that's what I read too. With low presence content signals the input "bridge" enhances the 3.5k band and as soon as there's enough energy (as measured by the comparator (the MC1433 that drives Q17) the vactrol dims the 3.5k band.
The discriminator seems to detect signal presence so in absence of signal the enhancement is cut.
That's just the result of a brief analysis before breakfast, so correct me if I'm wrong.
Looks to me like it's clearly dedicated to AM radio. It could probably be tweaked for producing an Aphex-type thing, but you would need two, and actually enhancer plug-ins cost nothing...

 

[NOON]

Can anyone with more discrete transistor experience than me help decode the sidechain sections? Even just breaking it up into functional blocks which I can then try and analyse individually?
Cheers!

 

[abbey road d enfer]

Here we go for a first round:
Q12 is a common-emitter stage with degeneration (the emitter resistor is not decoupled) so the gain is ca. the ratio of the collector resistor and the emitter resistor (veeeery roughly). With the actual values gain of ca.5 or 15dB.
Q13 is an emitter-follower, unity gain, low output impedance to drive the rest of the circuit, in particular the RLC filter R49/C16/L2.
Q14 is also a voltage follower, has a high input impedance so it doesn't damp the RLC circuit.
Q15 same as Q12 but gain about 20dB.
CR7 & 8 rectifiers.
Q16 voltage-follower, drives negative input of diff amplifier U2.

Simultaneously, voltage from input buffer (Q12/Q13) is rectified with CR10/11, which drive directly the positive input of diff amp U2.
Note that time constants of both rectifiers are similar, so any difference at the input of the diff amp will be due to amplitude differences between upper and lower path.

Resulting voltage at output of U2 goes via CR13 so only positive output is taken into account. Again a new set of time constants with R66, R67 & C27.
Q17 & 18 constitute a Darlington (very high current gain) used in voltage-follower mode, very high input Z very low output Z, coz drives light bulb - maybe not always, I don't understand the switch S2. Is it a 3-pos switch?

TBC

 

[NOON]

That's a great start, very much appreciated! All starting to make sense to me now.

The switch is indeed a 3-way, marked Speech / Music / Off. It's been tricky to trace out, it's a bit of a rat's nest and may still not be 100% correct, but it seems to be an ON/ON/ON rather than a centre off. It disconnects the CV in the off position, but in the Music position it seems to connect the simple voltage divider to the next stage input, vs connecting the previous Discriminator stages in the Speech position.

 

[moamps]

I will just add that PNP transistors should be flipped vertically in the schematic, IMO.

 

[abbey road d enfer]

That's a great start, very much appreciated! All starting to make sense to me now.

The switch is indeed a 3-way, marked Speech / Music / Off. It's been tricky to trace out, it's a bit of a rat's nest and may still not be 100% correct, but it seems to be an ON/ON/ON rather than a centre off. It disconnects the CV in the off position, but in the Music position it seems to connect the simple voltage divider to the next stage input, vs connecting the previous Discriminator stages in the Speech position.

Can you make that clearer by using a stack of 3-way switches?

 

[abbey road d enfer]

Q19/20 form a soft clipper with the string of diodes; low-level gain is R73/R71 (ca. 160), so any signal > 20mV is clipped to ca. 2.1Vpk. Actually simulation shows about 3Vp2p.

HOLD ON! Actually it doesn't fit.  The DC level after rectification is too low to turn on Q21. What are the diodes (18-23 and 24-27)?

 

[NOON]

Got time to do some more research. The diodes translate as 1n456A, a pretty standard small signal Si unit. there's a couple of zeners elsewhere in the circuit, but everything else looks like these.