ETI/Powertran/Elrad vocoder

[Magnus Danielson]

From: Hallvard Tangeraas <hall at oslo.online.no>
Subject: Re: ETI/Powertran/Elrad vocoder
Date: Sat, 17 Jun 2000 23:42:19 +0200

> At 18:02 +0200  12-06-2000, Magnus Danielson wrote:
> 
> >From: Hallvard Tangeraas <hall at oslo.online.no>
> 
> >> There's something weird with the circuits -as far as I can remember
> >> there was some confusion about the placement of the pots (there were
> >> some extra connection points, which made it very strange).
> >
> >Ah. Even if you did not say *which* pots, here is the secret:
> >
> >If you look really carefull you will see that the pots on the photos
> >have metal supporters. These supporters have their own holes and
> >solder points. Now, somebody has been smart and use these as ground
> >connection between parts of the PCB. Thus, if you look at the PCB each
> >pot have a hole frame like this:
> >
> >O       O
> >
> >O O O O O
> >
> >Where the 4 corner holes are the metal supporter connection and the
> >three middle onces is the actual pot ends and tapper in the middle.
> 
> I understand, but it's still confusing.

Well, I only told some things... but I have now digged deeper into
this so...

> I've had a friend scan the PCB/component placement as well as the
> circuit and plan to colour-code the connection points and upload to a
> web page, so we can take it from there.

Ah. This input board really needs a good workthrough it seems.

> Currently I'm having a hard time comparing the schematic with the
> actual PCB.

Join the club ;)

> I'm going through every component, and have covered a few
> connection points, but I'm still confused. I need some help.
> Do you have all the diagrams there?
> (it's a little hard to explain if you don't ;-)

Yeap! And I have it "in flesh", that is, I have the input boards
stuffed with components.

> OK, here we go....
> 
> 1) Looking at the schematic, there are some discrepancies...
> To the left side you have "IC2a" and "IC2b". This is the 1458.
> There are no pinouts, but I've looked the IC up, and looking
> at "IC2a", "+" is pin 3 and "-" is pin 2.
> The strange thing is that in the actual PCB, *pin 2* seems to go to the
> middle connector of "RV2" pot. What do you make of this?
> It should be pin 3 according to the schematic.

Since I have prepared myself with a blowup of the PCB layout and made
a new annotation of components it was a simple matter to verify
this. Yes, you are correct. I've checked the 1458 pinout in two
diffrent source and it has the same classical pinout. Also, in both
the PCB layout and my in-the-flesh PCBs pin 2 (- input) is really
wired to the tapper of the pot. So, if you follow the schematic you
would find that the PCB layout is incorrect.

However, if you look at the schematic and starts to ponder some over
it you would see that the - input is hooked to signal earth. Now,
hooking the - input to the signal earth would make that op-amp act as
a comparator due to its high open loop gain. Since it is intended to
be an EQ I don't think it would be wise to have a comparator at this
location. If we however assume that the PCB layout is correct and the
schematic is incorrect, then we would have the + input wired to signal
ground and the - input to the RV2 pot tapper. This would make
perfectly sence since now the RV2, C6, R9, R10, C7 and R11 form the
basis for a negative feedback network. This could probably work.

So, I'd say that the schematic is incorrect here. It is the only
sensible solution to this inconsistency.

> 2) The jacks...
> Looking at the schematic again, you have SK1 at the very left, then SK2
> a little to the right/top of it. Then there's a jack-switch a little to
> the
> right/under SK2.
> Is this part of SK1 or SK2?

It is part of SK2. You realize this if you read the "HOW IT WORKS *
Input Amplifier" text. The following sentence give it all away:

	If SK2 is used then IC1a is isolated and R4 is disconnected
        from ground making IC1b now having unity gain.

Looking at the schematic you realize that if SK2 is used means that a
plug is in SK2, then the IC1a is isolated means that this will break
the connection from the output of IC1a to something... Ah! See... to
RV1! So, when SK2 is used, then RV1 shall be the input divider for the
line input and if SK2 is not used, then RV1 becomes the input divider
for the MIC input, but after the initial amplifier. Hooking up the
wires to the PCB agrees with this.

> >On some PCBs there is explicit usage of the supporters for ground
> >connections so if you do not have the matching pots you are required
> >to wire these by hand.
> 
> I sort of figured that out last time, the biggest problem knowing where
> to connect the wires.

Yeap. This is totally undocumented for the connectors. You are bound
to make some reverse engineering here. However, these locations are
wisely distributed to match the position of the connectors and normal
location of that function on connectors.

The positions are:
IC1 side: Mic connector (SK1)
IC2 side: Line connector (SK2)
Near ICs: TIP & TIP Break
Near pots: BODY & BODY Break
Left of IC2 (solder side): LINE TIP Break and Line BODY Break
Right of IC2 (solder side): LINE TIP and LINE BODY
Left of IC1 (solder side): MIC TIP Break and MIC BODY
Right of IC1 (solder side): MIC TIP

Giving these general directions I think you should be able to find
the exact spots using some additional reverse engineering.

> >This thing confused me quite alot also, but when I discovered this
> >solution it all became clear to me. Let me tell you that it did cost
> >me quite some time to do the wiring.
> 
> Me too, and damaged PCBs from all the soldering/desoldering.
> This time I have fresh PCBs, and I'm going to figure out the
> connections *before* soldering!

Wise man. If you been really smart however, you would have figured it
out and redone the PCB Layout, but then naturally you fall into the
trapp that the documentation (with it's faults!) does not document
your PCB Layout but another, so you really have to document this new
layout better than the previous one.

> >Now, add this fact to the not allways so clear component placement
> >indications in the article and you have a good basis for confusion.
> 
> Agreed. It really is a bit too messy.

Right.

> >> And perhaps I should just get my equipment out of the closet and start
> >> building those PCBs again... any other suggestions?
> >
> >Ugh. I hope we are only speaking about the input boards.
> 
> Yes, fortunately :-)
> The other PCBs are OK.

Right. The input boards are done quite smart, but functionallity
suffers when the design does not properly carry through.

> >One confusing part is that some of the busses over the analysis board
> >is cut and should only operate over some of the channels. It took some
> >thinking to figure out how to do that. The trouble is, the info is
> >there, but it is a bit too confusing.
> 
> Hmmm.. I can't remember having any problems with that board. I
> completed it several years ago, but we can compare notes when I'm done
> with the input amps and I'm ready to connect everything together,
> hoping it won't blow up or anything!
> I spent 2 years or so finding the PCB layout for the main board!!!!!
> 
> ETI didn't have it (you were supposed to order it from Powertran, who
> designed the circuit and also sold the kit), but they had gone out of
> business, and ETI couldn't help me.
> Someone on the net (might even be on this list) directed me to the
> German magazine "Elrad", and a few weeks later I finally got the PCB
> layout!

Well, it is not very difficult, it is just that you have to be
carefull to understand what you are doing.

> >I for one has trouble with my input boards, I have somehow modified
> >one to function but naturally failed to duplicate this mod over to the
> >other board (they are equal designs!). Now, I can only blame myself
> >for not spending the quality time when my head is fresh enought, so
> >this is the main reason I haven't got my Vocoder operational!
> 
> But why not just look at the modified PCB and compare it with the
> other one?

Been there... however, the other nigth I found some connection that
was not correct. I don't recall what it was thought. Anyway, carefully
checking it up should make it apparent.

> I'm very interested in hearing about it.
> So you actually have one fully working input amp board then?

Yeap.

> >OK. I really should get it done and continue the trimming and setup job.
> 
> Same here. I've spent so much time and money on this project, and I'm
> still not done. I miss having a vocoder, after having sold my Roland
> VP-330, building this one to replace it.
> 
> >Oh, tip:
> >If you dont have a transformer with a separare 7.5V coil, but rather a
> >pair of 7.5/8 V tappers in the 15V coils
>   [.......]
> 
> I modified it slightly to accomodate for a different transformer I
> believe, but I can get back to this later. The input amps are enough
> for the moment ;-)
> 
> 
> >Another tip:
> >I got myself one of those standard plugs with builtin line
> >filter.
> 
> AC power plug you mean?

Yes, with builtin AC filter.

> >Anyway, what I think that this design should benefit from where a more
> >comprehensive description of how to build, how things functions, how
> >to wire things up etc. Add to that a rework of the schematics and
> >possibly also of the PCB and building it would go much smoother.
> >It is not that the articles are bad, but they where forced to compress
> >it into a small number of pages.
> 
> I think you're right.
> I've put a lot of work into this project, and along the way, asking for
> help I've instead received lots of requests about where to get the
> diagrams, more information etc. as they want to build one themselves!
> As far as I can remember, someone scanned the whole project and made it
> available on the web. If I succeed in completing mine I want to add
> information about these input amps and other things that can be
> misunderstood easily.

This someone is Anders Sponton. He also built most part of one, but he
sold it to me before he got to finish it. So I am sitting here with
the vocoder and the original copies that he scanned.

Cheers,
Magnus