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

[11-90-an]

Here is the schematic:

If there was a quota for questions asked per day i think i blew it by now...

Here goes...

Why is there no DC blocking cap in the buffer input stage? And could the JFET buffer be replaced with a BJT buffer?

And also, is there a specific reason why the 1uFs and 10uFs are ceramic here? Would there be a drastic difference with electrolytics?

[rankot]

Why is there no DC blocking cap in the buffer input stage? And could the JFET buffer be replaced with a BJT buffer?

Because there are caps later in signal chain which will stop DC passing further.

And also, is there a specific reason why the 1uFs and 10uFs are ceramic here? Would there be a drastic difference with electrolytics?

1u shall be polymer caps, not ceramic, as they're much better sounding compared to elco.
10u caps are in switching module, so using ceramic will have lower ESR and smaller footprint.

[jonny.reckless]

As Ranko said, you get less losses and better HF performance by using film and multilayer ceramic caps rather than electrolytics. They've gotten a lot cheaper in the last few years so you might as well use ceramic for 10uF, and electrolytic for bigger values say 47uF upwards. Also in some circuits you want to block DC. Electrolytics don't like having zero bias across them or being reverse biased, it messes with the chemistry of the electrolyte; you don't need to worry about that with ceramics so they make better DC blocking caps for audio circuits generally.

[PRR]

> Why is there no DC blocking cap in the buffer input stage?

What DC bias voltage is needed at this JFET's Gate?

> And could the JFET buffer be replaced with a BJT buffer?

What DC bias voltage is needed at this BJT's Base?

[kgusarov]

Let me know if you want me to share the Gerbers or NC drill file for the 8 knob redux version.

Definitely! Only thing, I am planning to make it as a rack unit, so... I guess, I won't need 7660 But this is something, I can handle myself. Sadly, not true for good PCB layout

[jonny.reckless]

Let me know if you want me to share the Gerbers or NC drill file for the 8 knob redux version.

Definitely! Only thing, I am planning to make it as a rack unit, so... I guess, I won't need 7660 But this is something, I can handle myself. Sadly, not true for good PCB layout

Here you go:
https://drive.google.com/drive/folders/1usNiADluT3tfycfkl7Y7Yo24fzZkVLZN?usp=sharing

[kgusarov]

Here you go:
https://drive.google.com/drive/folders/1usNiADluT3tfycfkl7Y7Yo24fzZkVLZN?usp=sharing

Thank's a lot! Also - brilliant work on the noise gate. This is something I am currently planning to tinker around. So far, I've came up with idea to level shift gate's control voltage from current implementation and then use it to control the THAT2180 VCA. Only problem here is, that I need a reference voltage to be set using a trimmer (VR4) on schematic. Maybe someone has better ideas? )

[rutgerv]

Hi kgusarov,

Interesting indeed! I'm actually in the process of transforming the ET to work with an SSI2164 VCA. In many ways this vca is similar to the that2180. Keep in mind that the original ET uses a LM13700 with lineair control law, while the abovementioned VCA's are expotential in their response to the control signal. For gating the control law may not be so important, but for compression it is. I'm using a second VCA inside the loop of an opamp to transform the ET's lineair control signal into what the expo VCA likes to receive. Will post with progress update soon.

Rutger

Rutger

[kgusarov]

I'm actually in the process of transforming the ET to work with an SSI2164 VCA.

Well, AFAIK, Analog Devices mark this as obsolete, and I don't trust any other manufacturers (this is my thing of trusting official ones only). So, I'd stick with THAT.

[PRR]

> Analog Devices mark this as obsolete

That's the SSM parts. Rutger is saying SSI. Same function under yet another company. I remember these parts from the 1980s. They go in and out production here and there.

Dan Parks is CEO of SSM and former Audio Products Director at Analog Devices and National Semiconductor.
https://audioxpress.com/news/sound-semiconductor-expands-ic-range-with-new-single-and-dual-vca-chips
https://www.soundonsound.com/news/sound-semiconductor-release-huge-quad-vca-data-sheet
https://en.wikipedia.org/wiki/CEM_and_SSM_chips_in_synthesizers

[kgusarov]

Same function under yet another company.

Still, this is my thing You can call me snob, but I trust TI, Analog and other big companies. I know, others do good things, but this is why I buy Samsung instead of Xiaomi Way of living, I guess ) Also, I buy from Mouser/Farnell due to my living place... And haven't seen small brands there...

[PRR]

SSI is a fully kosher company with roots in this technology.

Yes, it is a >teeny< company with products that went "obsolete" (insufficient demand) at 2 or 3 other companies already.

And they only sell through a few distributors, also small shops not Mouser/Farnell. I'm sure there are places and people who just should not try to buy SSI.

Much depends on your production. For a solo guitarist with good delivery service, you can just buy a 10-pack which will be a lifetime supply. (What I did when the LM377 chip was shut-down: I bought a handful and they lasted until I left that place and quit abusing that amplifier.) If you are starting a factory and signing with Banjo-World to supply a million ETs a year, you want to avoid a small unsure supplier (or else buy them as a pet).

[kgusarov]

And they only sell through a few distributors, also small shops not Mouser/Farnell

This already goes to the offtopic But haven't found any trusty place in EU...

To stay more focused on the topic itself... I was also thinking on using JFET, but that seems easier to me so far. Since I can provide kind of reference voltage to EC+/EC- and scale the gate CV so, it will definitely be closed. When EC+ = EC- gain is 0db, which is great point for gate openning. Only problem atm for me is that I am out of ideas on how to generate stable VREF without need for a precision trimmer and millivoltmeter...

[rutgerv]

I usually use the Lm4040 for reference voltages.

[kgusarov]

I usually use the Lm4040 for reference voltages.

I usually prefer LT1009 or TL431. However, it still would require a lot of tuning with high precision instruments due to high sensitivity of THAT IC. Even slight difference will cause the signal to be boosted/attenuated. On the other hand, it should be possible to use JFET still, I guess, or even put LM13700 in signal path for "coloring".

[jonny.reckless]

If you're using a THAT VCA I recommend one of their excellent application note designs as a starting point. I've built a couple and they are studio quality and sound fabulous on a wide variety of program material. It's important to get the power supplies, grounding and impedances right to get the best low noise low distortion performance out of those chips. In particular the control port needs to be driven with a low impedance, low noise, low distortion amplifier. You don't want any TL072s anywhere near it.

http://www.thatcorp.com/datashts/dn01A.pdf
http://www.thatcorp.com/datashts/dn00A.pdf
http://www.thatcorp.com/datashts/dn107.pdf
http://www.thatcorp.com/datashts/THAT_4305_Datasheet.pdf

The engineer's thumb is a completely different beast and I fear trying to combine it with a THAT VCA you will end up with an expensive mule, i.e. neither one thing nor the other, since the CCA sidechain law is linear in the ET and exponential in the THAT for a start.  The ET is a great low cost simple compressor which excels for what it was designed.

[kgusarov]

I agree, that it will be step away, but it is always nice to do some experimentation. TL072's are there due to the fact, that they ended up faster to simulate in LTSpice, than other opamps So, I put them in every design, since they converge so good. I was basically thinking of turning this into 2-stage thing - first gate, then compressor, otherwise it seems, that compression ratio also affects gate response.

[jhergonz]

Good day guys, how are you?

Anyone here tried building the latest version of Engr's thumb?

I tried building the 2019 version of Engr's thumb, it is working but has several issues,

instead of having compression, the signal is being muted, then the sound will comes back.

and when i turn the RATIO control on, there is no sound coming out when i reach 9o'clock in the RATIO control.

i checked everything and found no shorted trace. double checked the components and everything is in right values.

Any one has idea what is going on with that circuit?

[rankot]

Anyone here tried building the latest version of Engr's thumb?
I tried building the 2019 version of Engr's thumb, it is working but has several issues,

Did you just built it using schematic, or one of the available layouts (which one)?

[jhergonz]

Anyone here tried building the latest version of Engr's thumb?
I tried building the 2019 version of Engr's thumb, it is working but has several issues,

Did you just built it using schematic, or one of the available layouts (which one)?

just the schematics