Neve Output Stage Addition to Console

[Greg]

It would a straight replacement, but at first glance, I wouldn't have to do any heavy modifying to the console PCBs. Let's assume, for the sake of argument, that I replace IC4a with a discrete opamp. Just mount the opamp on small perfboard and connect it up. If I change the entire IC4b stage with the 283 AM stage, I THINK I can take the wiper of the master fader right to input of the AM stage.

I don't really see the point in making something switchable... at least not in this case.

 

[PRR]

> isolate the bus-parasitics from the inverting input of IC4-pin#2.

Seems likely. It might just be a jumper.

> Regarding IC4a, I do have a couple MM-990

A true 990 would be an obvious, though costly, fit here.

5532s are not all the same. You might like JRC better than some other recent batches.

> the 220R is not setting gain

Correct.

> nor is the 47K going back to the inverting input.

It sets the I/V conversion factor. 1mA in gives 47V out. Or to a simple ratiometric mind: the gain from one channel panpot is 47K/22K. If you are going to do some type of virtual-earth mixing, and I think you are forced to unless you want to rip-out most of the mixing and assignment system, then your mix-amp will probably be an inverter with about 47K feedback.

> IC4b is an inverting gain stage, right? A=(22k+8K2)/12k=2.5 ?

Looks that way to me. Don't forget that a balanced output will be twice as much, or 1:5 from the master pot wiper. You might be able to run the 283 second stage at such low gain. Or at a higher gain with a voltage divider somewhere in front.

 

[clintrubber]

[quote author="PRR"]> isolate the bus-parasitics from the inverting input of IC4-pin#2.
Seems likely. It might just be a jumper.
[/quote]

Must say I was surprised to see added resistance exactly there, since that point likely better be the most virtual earth it could be etc, but (as I think to understand it at this time of the night) it doesn't spoil the summing-performance.

Regards,

Peter

 

[PRR]

> it doesn't spoil the summing-performance.

220 ohms in series with the input pin impedance, which is surely over 200K, does nothing on paper.

In real life, there are parasitic capacitors everywhere. I would expect the 220 to be outside the loop, but maybe layout likes this way better, maybe the drawing is wrong.

 

[Greg]

[quote author="PRR"]... maybe the drawing is wrong.[/quote]
I'll do a quick trace and see what I come up with.

Regarding, the MM-990, I have 2 of these:
http://www.mil-media.com/docs/products/mm99.shtml

I don't know if these are the originals designed by Fred or if they've been modified. Regardless, I'm sure it'll kick the crap out of a 5532.

 

[PRR]

> I have 2 of these:

"...format-compatible with Jensen / Hardy 990tm ... unique, all-discrete Field Effect Transistor signal path."

See if Fred will jump in here. There is, I think, one "990 pin compatible" with very different DC offset specs, such that you want to consider and maybe revise the DC gain and bias. Since in any case, you will be cramming 2 inches of stuff in a half-inch space, the DC affair is a small detail, but if you omit it (and if these are not zero ofset) then the amp will swing to a supply rail and be stuck.

Oh: when hacking a 2" module into a 0.5" space, you can't run long sloppy wires. Mix-amps are fussy things to begin with. Simply for that reason, I'd be inclined to explore fancy caps and alternate chips before going to a large modular amp.

An AC-coupled summing amp does not have to be a full OpAmp. A 2-transistor stage would work, with less precision but less complicated errors than a chip.

 

[Greg]

I did some work on my output stage this evening. I first put a socket for IC4 since there are NO sockets anywhere in the Saber. Before getting too crazy, I decided to simply replace IC4 with a OPA2604.. ie IC4a and IC4b were replaced. There's another 5532 used for balancing after this stage which I left alone. The results were good, but nothing earth-shattering to me.

I proceeded to jig up a discrete opamp to replace IC4a but still use half the 5532 for IC4b. After some fiddling and one failed attempt, I successfully got the left channel of my console working with a Melcor 1731 for IC4a. Since I used sockets and pins, I swapped the Melcor 1731 for a MM-990, and this is when the differences became more noticable. The midrange difference between the left (modded) and right (stock) channels became more evident.

However, the most noticable difference tonight was replacing IC4a with a MM-990 and IC4b with half the OPA2604. On comparison, the stock channel sounded grainier... most noticable by guitar tones.

I don't want to jump to any conclusions just yet, but I feel I'm making some progress. I'm getting the few parts I need to complete the 283AM, which I'll use to replace everything after the master fader. I'm hoping this change will add some "muscle" to my mix (so to speak).

Very cool !!! :thumb:

 

[longsoughtfor]

Way to go Greg! This may require pictures though Do the discrete modules fit between the two channel boards OK? I'd be interested to see that.

I have to disagree with PRR slightly, I do think upgrading the master section is a decent thing to do with this board... for $200 it's hardly overkill and if the improvment is noticable then it's well worth the time too. However, I also think that figuring out the grounding and power issues may make just as big an improvement to the 2-bus sound quality.

Cheers
Kevin.

 

[Greg]

The discrete modules don't fit worth a crap. It's a total hack job at the moment. It's really cramped in there, but it definitely works and is an improvement. I'm still waiting on some parts for the Neve out. Once I get that working I'll probably figure out a better way to implement the discrete modules. I might end up moving the entire IC4a stage to another PCB. Have to think about it more, though.

 

[Greg]

The parts I ordered over a week ago still aren't here... My box that has the rest of my Neve AM stage been sitting in Texas for almost a week now. Shipping isn't the greatest around here because of Katrina. She won't stop haunting me :evil: :twisted: !

 

[Svart]

shielded wiring to/from the module? i don't know if that would help or not or if the problem would really be related to wire length rather than noise coupling..

either way I am considering taking up the job of laying out an Ideal buss channel using these discrete opamps..

 

[Greg]

I finally receieved the parts I needed to finish this Neve stage. It only took two and half weeks to get something from Texas. :evil:

You should be hearing back in a week or so, hopefully with some good results :wink: .

 

[Greg]

Here's a funny story for you:

I finished stuffing the Neve output stage tonight, and decided to fire it up on my bench... and guess what ?!? NO OUTPUT.

Crap...

So I'm going through the basic things of checking for wrong resistor values and short and things of that nature. An hour and a half later everything looks fine and still no output. Smoke break time !!!

I come back and decide to check DC bias. I check Q5 everything looks fine. I check the collector of Q4 and all is happy, but then I check the base and emitter and, uhoh, zero DC bias. Ok here's the problem. So I start checking around Q4 and still can't find the problem. Dammit !!

So I decide to see what's going on with the 2N3055. I'm not that familiar with this transistor so I fire up the datasheet. On the first page of the datasheet is states: CASE IS COLLECTOR.

SHIIIIIIT !!! that's my problem... when I mounted the damn heatsink to the 2N3055 I used nylon insulators, so the collector was basically floating. So take off the nylon insulators so the mounting screw is making contact with the case and the solder pad, fire up the Neve and now I see my output... clean as can be. I feel like an idiot, but I guess I'm glad I found the problem on my own, without having to bug you guys... it only took me 3 hours.... sheesh.

I'm going to have a beer now, and maybe this weekend I'll try and put this sucker in my console.

:thumb:

 

[Bluzzi]

Greg, this is one of the most interesting threads I have been following! I wanted to do this with my console but with Hardy 990s. I haven't had the time, money or mainly guts to take on such a project.

I am waiting in anticipation of your success.

Jim

 

[Greg]

Thanks for the encouragment Bluzzi :thumb: !

This is by far the gutsiest project I have undertaken. I've been trying to think of what I still need to do, and here's me thinking out loud:

1. Match gain of this Neve out stage with IC4b stage.

2. IC4b, which I'm replacing with the Neve out, is an inverting stage. IIRC, the Neve out is not inverting. So I'm thinking I need to flip the signal back. Can I do this by swapping either the primary or secondaries on the output tranny?

3. Grouding - I'm using a seperate PSU for the +24VDC, so I'll need to avoid ground loops. I'll need to look at the Saber grounding scheme.

Can anyone see anything else I need to consider.

 

[Family Hoof]

[quote author="Greg"]Neve out is not inverting. So I'm thinking I need to flip the signal back. Can I do this by swapping either the primary or secondaries on the output tranny?[/quote]
Just swap the secondary leads (i.e. XLR pin 3 hot instead of pin 2). The transformer won't care. No harm done.

 

[Viitalahde]

[quote author="Greg"]I'll need to look at the Saber grounding scheme.[/quote]

Just take your ground lead and stick it around 'till it stops buzzing. :twisted:

 

[Svart]

realistically you want to source your input AC from the same outlet as the console at least and your chassis potentials need to be equal. grounding through ferrite might be worth a look but I don't think you'll have too much problem as long as the current flowing over ground isn't excessive.

Keep us informed!

 

[Greg]

Thanks Jens! Nice and simple. :thumb:

I don't expect the grounding to be a huge issue. I'll just take an educated guess, and if that doesn't work I'll just poke around as Viitalahde suggested. :grin:

 

[Greg]

I took a closer look at the schematic last night since I'm getting ready to give this Neve output a whirl. I was hoping to get it working over Thanksgiving, but that plan didn't work out. Anyway, upon first glance, I assumed that IC3 was feeding the unbalanced 300mV output. But IC3 is also feeding the "monitor PCB" which feeds my control room outputs, among other things. So, I do indeed need the IC3 stage. I probed the "unbalanced" output on the Neve out last night, and the signal look great. I'll attempt to feed the monitor PCB with this... what do you think Jens?