THE ENGINEER'S THUMB... At last, a better compressor!

[Kevin Mitchell]

After comparing the strip board layout with CA3080 I've noticed the following;

- The 2n9001 has been omitted. Is this a bad thing?

- the 1k and 10n on the input has been switched around. Don't think that's a big deal. Maybe has to do with not using the protection LEDs.

- Millennium bypass variant has been omitted.

After comparing the ICs (CA3080 and LM13700) I've drawn this up and believe that the connections are fine.


So why are people still using the LM13700? Does it come down to price and availability?

[italianguy63]

The LM13700 is like $1.10

vs. 2 x $2 for CA3080's... that's the way I see it anyway.

MC

[armdnrdy]

The LM13700 is like $1.10

vs. 2 x $2 for CA3080's... that's the way I see it anyway.

MC

The way that I see it is....the CA3080 has become a target for relabeling.
There are a lot of counterfeit 3080s running around out there.
Check the pinout of a single op amp to the 3080. Your circuit will pass signal, seem to be working but...it's actually not.
The suppliers that have actual 3080s have raised the prices dramatically.
I have yet to hear of a relabeled 13600/13700.

Edit:
You might find this useful.
http://www.diystompboxes.com/smfforum/index.php?topic=105341.0

[Kevin Mitchell]

I'm definitely going to draw up and test out this Engineer's Thumb with a CA3080 and a TLO74. More trace work designing probably but should take up less board space. I do plan on adding attack, release and treble boost switch to my project. Not threshold though - Merlin states a few times it functions the same as Ratio. Then I'll have Groovenut beat for design 

I'm only teasing. Beautiful work by the way, Groovenut. Very impressive. Nice site as well.

The way that I see it is....the CA3080 has become a target for relabeling.
There are a lot of counterfeit 3080s running around out there.

Wow I didn't know that was a thing - counterfeiting ICs lol...
I was going to purchase at least one from smallbear I'll try contacting Steve to see if he has anything to say.

[italianguy63]

The "counterfeit" CA3080's I was getting were OK until about 6 months ago.

Now they sound like CRAP.  Noisy, unusable.

I have tried about 3 or 4 different sources.  All bad.  I posted about this a couple months ago.  You can get NOS ones for about $5-$6 each.  But, they are just priced out of the market at that level.

[armdnrdy]

Wow I didn't know that was a thing - counterfeiting ICs lol...
I was going to purchase at least one from smallbear I'll try contacting Steve to see if he has anything to say.

Steve's will be good. He's privy to the scams. I'm sure he tests them.

[italianguy63]

Wow I didn't know that was a thing - counterfeiting ICs lol...
I was going to purchase at least one from smallbear I'll try contacting Steve to see if he has anything to say.

Steve's will be good. He's privy to the scams. I'm sure he tests them.

Yes.  His will be good NOS pieces.  No problem there!

MC

[italianguy63]

Another metal can NOS equivalent is NTE902.

MC

[armdnrdy]

Another metal can NOS equivalent is NTE902.

MC

Yeah but....that part seems to average about $18+ 

[italianguy63]

Another metal can NOS equivalent is NTE902.

MC

Yeah but....that part seems to average about $18+ 

Just sayin'!

I've stumbled across a few cheaply.. 

[PRR]

Metal vs plastic: it is the SAME chip inside.

When chips cost $50 each, and went mostly to the military, it made sense to use metal cans and proven (tube-era) metal-glass seals. But plastic is so much cheaper, and in the 1960s TI was making many million plastic transistors for commercial computers with excellent track record. Old habits linger and so did metal-cans, but you absolutely can use one for the other in any on-earth application.

"Counterfeit" is often not a work-the-same chip but a work-"sorta" chip with the factory markings washed off and new marks applied. Like if I got a KIA and pried-off all the bages, put Acura badges on. An innocent buyer might not know the difference- it starts it runs etc. Except Kia/Acura, the fenders are different; most chips are all-the-same except the badging. Parts-marking machinery is very readily available in any electronics neighborhood (mostly China these days). When '3080 went out of production and retail prices rose, some clever sneaks saw an opportunity. Especially since many '3080 builders do not build right away, or are newbies who won't be able to trace their failures to a bogus chip. And there are a few '3080 uses where an op-amp will "work OK"-- just not in any gain-control application. So the eBay hawks are betting they can buy 10-cent chips, mark them for 1 cent, sell them for a dollar, and come out ahead even if some buyers get wise and complain.

I've seen several builds here, even from Known Name kit-stores, which didn't work until a GOOD '3080 was tried.

> His will be good NOS pieces.

AFAIK, Steve is preparing to transition from verifiable odd-lot NOS to new-made on original masks and low-tech processing. ('3080 was low-tech when new and is very training-edge in today's semiconductor word). There are operations like Rochester who specialize in keeping old-old chips in-stock as global demand falls from millions to thousands to hundreds, and even new-make them 1,000 at a time if they think they will sell them over a few years. Steve of course has a very clear idea of the pedal-parts market for '3080s and can encourage or front-money Rochester if that is what has to happen.

[italianguy63]

Metal vs plastic: it is the SAME chip inside.

Exactly, Paul.  No problem getting the plastic "counterfiet" Chinese (good/working) versions of the DIP CA3080 for $1.50 to $2.00 if you shop.  I've just quit fighting it, and have given up on the metal (CAN) versions for now.

I still think, in this context, why use two ($2x2) CA3080's when just one 1st grade (Mouser) ($1.50) LM13700 will suffice?!



A bunch of NFG CA3080 CANs.  A couple DIPs and a NTE902 looking on:

I've wasted a lot of my time on this...


MC

[samhay]

I'm thinking since I'm into more subtle compression the increase in threshold should  get me where I want. Question though, increasing the 1k on the emitter would reduce current to the OTA therefore decreasing the output of the OTA via control current whereas increasing the 10k at the + input of the OTA would reduce the output via less input, both resulting in more overall output level, however will the method used change the outcome or is it 6 to 1, half dozen to the other?

There is somewhat a redundant cycle in reducing the input at + on the OTA but I would think reducing the current drive, being outside of the feedback loop, would be more direct in affect.

I would imagine from a signal-to-noise perspective, reducing the current into the bias pin would be prefereable. Also, I suspect Merlin spent a good while deciding on the 10k/220R divider into the op-amp to maximise the clean headroom.

As far as CA3080 vs LM13700.
They are not entirely equivalent, but are interchangable in some designs including the ET. Personally, I would keep the 3080 for older designs that rely on mojo, like a Ross/Dynacomp. Also, the layout may not be any smaller using a 3080, as you can push the 13700 to the edge of the board and route everything to one side.

[Kevin Mitchell]

As far as CA3080 vs LM13700.
They are not entirely equivalent, but are interchangable in some designs including the ET. Personally, I would keep the 3080 for older designs that rely on mojo, like a Ross/Dynacomp. Also, the layout may not be any smaller using a 3080, as you can push the 13700 to the edge of the board and route everything to one side.

If they are interchangeable without any discrepancies for the ET then it would make sense to use the smaller IC. I can understand how using the smaller version of the LM13700 as most people in this thread has is an ideal option but lets say someone with some sort of "healthy" OCD -lol. For me it's about consistency and eye appeal even though the circuit is hidden in a box. If I can get away using a CA3080 (8 pin) and a TL074 (14 pin) I see that being a great/practical choice. I'll report back with my findings soon.

[armdnrdy]

As far as CA3080 vs LM13700.
They are not entirely equivalent, but are interchangable in some designs including the ET.

The LM13700 is nothing more than a dual 3080 with a darlington buffer and linearizing diodes in a DIP16 package. If you do not connect the buffer and diodes...you have a 3080.

The design story here:
http://www.idea2ic.com/LM13700.html

Now if you were to sub two 3080s for a 13600/700, you would need to check the connections in the circuit and add a buffer and/or linearizing diodes if those are being used.

[samhay]

The LM13700 is nothing more than a dual 3080 with a darlington buffer and linearizing diodes in a DIP16 package. If you do not connect the buffer and diodes...you have a 3080.

This is mostly true, but take a look at the data sheets. The amplifier bias is set up differently such that this pin sits 1 diode drop above V- for the 3080 and 2 diode drops for the 13700. This makes a difference in some applications, but the ET uses a fairly sophisticated precision current source to drive the amp bias, which makes it immune to the voltage on the bias pin. Also, while the ET doesn't use the linearising diodes (not sure how you would add these externally) and output buffer, many applications do.

[samhay]

If they are interchangeable without any discrepancies for the ET then it would make sense to use the smaller IC.

It would make sense to you. Not so much to some others.

[armdnrdy]

You are correct Sam.

Looking at the 3080s amp bias input, the diodes anode is connected to the base of the input transistor....where as the 13600/700 is connected to the emitter making for two diodes in series.

Adding linearizing diodes to a 3080 would be fairly simple.
The cathode of one diode is connected to the inverting input.
The cathode of a second diode is connected to the non-inverting input.
The two anodes are connected together making the diode bias input as in the 13600/700.

Now...I don't know why anyone would do that when you could just use a 13600/700 except in tight layouts.
While researching the 13600/700 to 3080 daughter board...I did come across a 3080 to 13600/700 daughter board. I guess the individual that worked up that conversion had a truckload of 3080s! 

[Kevin Mitchell]

I've compared this daughter board with Merlin's 1st layout(2nd post) which is why I drew the IC schematic on top of this page. The conversion is simple when using for the Engineer's Thumb.

If they are interchangeable without any discrepancies for the ET then it would make sense to use the smaller IC.

It would make sense to you. Not so much to some others.

Regarding the Engineer's Thumb I can not agree with your reply. For this project they seem to work 100% the same with 100% the same results. For other applications I however do understand your point. I've said "interchangeable without any discrepancies". If I'm missing something please elaborate.

I'd rather not work on an inferior compressor which is why I've been barking up this tree. Also - 2 of 8 pins unused apposed to 10 of 16. I sense some practicality here.

That's all I'm getting at.

[samhay]

>For this project they seem to work 100% the same with 100% the same results.

I don't think anybody has said that, but rather that the 3080 can be used - I would imagine the noise and overload characteristics are not 100% identical.
This might seem overly pedantic, and perhaps it is. In any case, go ahead and use a 3080. I bought a bunch of these some years ago and am waiting for the price to go up before selling them off - the more demand for them the better.


Larry - your probably right about using external diodes, but you won't get thermal compensation, which probably doesn't matter. I wonder if anybody has tried this with a 3080 or similar.