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

[iampoor]

So Ive read through the whole thread....and Im still having some issues.With any compression, Im getting some noticable distortion. It sounds like parasitic oscillation/instability. On lower ratio settings, it starts distorting that sort of "fizzes" out. I have built the circuit a few times, and am using genuine LM13700's (Got them from TI).

VOltages seem fine, Ive done multiple builds and the schematic/builds seem to match. Any idea where to start from here?
Im writing this at 3am.....very discouraged, so sorry if Ive missed something obvious!

[iampoor]

https://youtu.be/OEzybiIzO-4

Little video showing the issue

[Cozybuilder]

Are you getting any signal on the collector of the transistor? It kind of looks like you are getting just amplified signal out of U1a feeding the output. Might as well check and see if the 1K from the transconductance out is really 1K, not infinite.

[iampoor]

Are you getting any signal on the collector of the transistor? It kind of looks like you are getting just amplified signal out of U1a feeding the output. Might as well check and see if the 1K from the transconductance out is really 1K, not infinite.

Actually, I am only getting around 1v DC on the collector, and I cannot see any AC signals on my oscillioscope....hmmm

Looking at the signal, I dont see much of anything except some small ac bursts, that start at the output of u1b. WHat should I be seeing on the collector?

[PRR]

> 1v DC on the collector, and I cannot see any AC

This is correct; and this is NOT a good point to be poking at. It should always sit very near 0.6V (or 1.2V?), and is VERY sensitive to static discharge.

The OUTput of the OTA (goes to the 1K which Russ mentioned, into the signal opamp) should show signal for large inputs. Also look at the emitter of the transistor-- this sits high at idle and drops negative for large signals.

[Cozybuilder]

I think the problem is related to the TCA- either its not getting input, or output, or the 1K is open, or the TCA is defective. You've checked voltages, and component values. From your video, it looks like the first op-amp is simply amplifying and clipping the signal as you increase the feedback (the 1M pot). So I would concentrate initially on verifying signal into the TCA, and the bits surrounding the TCA (wiring, solder joints etc.) It could simply be a bad chip too.

[sajy_ho]

Hi guys, I built the circuit using Merlin's PCB and having exactly the same issue as once mentioned here:

https://youtu.be/OEzybiIzO-4

Little video showing the issue

Following your above posts I checked all the connections around the input and output of the OTA and everything seems to be OK.
Also I replaced the chip with another LM13700, and even subsituted it with a CA3080 and still having the same problem. Here is my voltage readings:

Q1:
Pin 1:  1.03
..... 2:   4.86
..... 3:   4.5

U1:
Pin 1:   4.5
..... 2:   4.5
..... 3:   2.21
..... 4:   0
..... 5:   2.21
..... 6:   4.06
..... 7:   4.5
..... 8:   9.02

U2:
Pin 1:   4.86
..... 2:   4.5
..... 3:   4.5
..... 4:   0
..... 5:   3.00
..... 6:   4.5
..... 7:   4.5
..... 8:   9.02

U3:
Pin 1:  1.03
..... 2:   0
..... 3:   4.5
..... 4:   4.5
..... 5:   4.5
pins 6-10:   0
pin 11:  9.02
Pins 12-16:  0

So what could cause the OTA not work?
Any help would be appreciated...

Thanks
Sajad

[Groovenut]

Just wanted to say thanks to Merlin for the great circuit! I had to do a few tweaks to get it to my spec but that's because I love subtlety and as speced was far from subtle ( I realize that's part of the point, a Dynacomp is far from subtle too). I do love that this has the ability to be as subtle as my optical comps and yet can get rude about it too. 







Note: these pics were taken before my subtlety mods and are of Merlin's original posted values

[Groovenut]

Quick question for Merlin,
Is there any way to get a slight overall volume increase? It would be nice to have a bit more room to go at the end of the level pot

Thanks!

[samhay]

I had to do a few tweaks to get it to my spec but that's because I love subtlety and as speced was far from subtle....

Care to share what your tweaks are?

Oh, and a very nice build!

[Kevin Mitchell]

Quick question for Merlin,
Is there any way to get a slight overall volume increase? It would be nice to have a bit more room to go at the end of the level pot

Thanks!

I'm not Merlin but I've read that increasing the voltage (some have experimented with charge pumps) provides a noticeable gain in headroom.
I have a couple questions myself if anyone cares to answer.

Is a LM13700 suggested because of it's price and availability "~$1.10+"? There's a strip board layout using a CA3080(an 8 pin IC used in a Dynacomp I've read?) "~$5+" I've read all over the place that the LM13700 is two of the CA3080s. With some obvious tweaks I'm not in tune with it yet. I'll draw out it's schematic to compare later.

First off I've noticed IvIark -the guy who shared it has omitted the 1n4001 as well as the BS170 MOSFET. Are those in Merlin's schematic for some kind of "old-school" bypassing, for use of an VU meter on the LED of the MOSFET or some other reason?

Random question out of curiosity, I believe there are TO-99 (metal can type package) of CA3080 "ua308HC" as well as TO-99 dual opamp ICs such as "LM358H". Still trying to find a decent match for a TL07x type IC. Although not ideal due to the price - is it possible/practical to use the metal can type of ICs instead?

I'm still a young learner trying to wrap my head around it all. Thanks for everything guys.

[bluebunny]

First off I've noticed IvIark -the guy who shared it has omitted the 1n4001 as well as the BS170 MOSFET. Are those in Merlin's schematic for some kind of "old-school" bypassing, for use of an VU meter on the LED of the MOSFET or some other reason?

It's a Millenium bypass variant so that you can switch the both the LED and the effect on and off using only a DPDT.  (Check out GEOFEX for more on this - link at top of this page.)

[italianguy63]

There is a metal can CA3080-- I will advise to steer away.  I have gotten them in the past from China to keep the cost down, but the quality has nose-dived in the last year, and I CAN NOT find a satisfactory source anymore.  (Other than expensive NOS supply).

The plastic DIP style CA3080A is cheaper/easier and readily available, however.

MC

[Kevin Mitchell]

It's a Millenium bypass variant

That's what I assumed but thought it was odd seeing this in a schematic dated in 2012. I refreshed my knowledge on everything regarding the millennium bypass last night. Using it is redundant these days due to the low cost of 3pdt switches. Not sure if the 2n9001 is part of the bypass as well so I'm hoping someone could explain.

Also is there any differences on the effectiveness of the circuit between using a LM13700 or a CA3080? I'll snag both to breadboard next week.

There is a metal can CA3080-- I will advise to steer away.  I have gotten them in the past from China to keep the cost down, but the quality has nose-dived in the last year, and I CAN NOT find a satisfactory source anymore.  (Other than expensive NOS supply).

The plastic DIP style CA3080A is cheaper/easier and readily available, however.

MC

Thanks for the info! I've noticed smallbear had a ua308HC and thought it would be cool to use the TO-99 type ICs which seems uncomon. I'll steer away and try to keep this hobby budget friendly 

[Groovenut]

I had to do a few tweaks to get it to my spec but that's because I love subtlety and as speced was far from subtle....

Care to share what your tweaks are?

Oh, and a very nice build!

thanks Sam!

Realistically all my current tweaks can be easily done on the 5 knob version but since 5 knobs won't fit here's the list.

Reduced release time, 470k
Increased attack time, changed 100R to 47R and a series 100k trimmer
Changed the Ratio pot value to A500k since I found the most aggressive compression I use was just past noon on the original value.
Slightly raised the threshold to accommodate higher output pickups with more subtlety.

I'm still thinking there may be a way to gain a little more output without upsetting the balance of the circuit though. Still experimenting.

[Groovenut]

Quick question for Merlin,
Is there any way to get a slight overall volume increase? It would be nice to have a bit more room to go at the end of the level pot

Thanks!

I'm not Merlin but I've read that increasing the voltage (some have experimented with charge pumps) provides a noticeable gain in headroom.
I have a couple questions myself if anyone cares to answer.

Merlin has said the circuit is optimized for 9 volts and increasing the supply won't actually help much. I'm definitely not having headroom or distortion issues. It's a very clean circuit. I would just like to afford some more room between unity gain and the end of travel on the level pot

[samhay]

Reduced release time, 470k
Increased attack time, changed 100R to 47R and a series 100k trimmer
Changed the Ratio pot value to A500k since I found the most aggressive compression I use was just past noon on the original value.
Slightly raised the threshold to accommodate higher output pickups with more subtlety.

I'm still thinking there may be a way to gain a little more output without upsetting the balance of the circuit though. Still experimenting.

Thanks.

If you want more output without further increasing the threshold - I take it you used a 1M trimmer like the 5 knob version? - then you could try increasing (don't decrease!) the value of either:

1. the 10k resistor feeding the LM13700's (+) input.
2. the 1k resistor between Vref and the PNP's emitter.

[Groovenut]

Reduced release time, 470k
Increased attack time, changed 100R to 47R and a series 100k trimmer
Changed the Ratio pot value to A500k since I found the most aggressive compression I use was just past noon on the original value.
Slightly raised the threshold to accommodate higher output pickups with more subtlety.

I'm still thinking there may be a way to gain a little more output without upsetting the balance of the circuit though. Still experimenting.

Thanks.

If you want more output without further increasing the threshold - I take it you used a 1M trimmer like the 5 knob version? - then you could try increasing (don't decrease!) the value of either:

1. the 10k resistor feeding the LM13700's (+) input.
2. the 1k resistor between Vref and the PNP's emitter.

thanks Sam. That was my thought but wanted confirmation. 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.

[bluebunny]

Not sure if the 2n9001 is part of the bypass as well so I'm hoping someone could explain.

2N9001?

Just this bit is the bypass:



(Along with the two halves of the DPDT switch, of course.)

[Cozybuilder]

The Millennium bypass needs the LED and load resistor on the drain to function reliably.