Author Topic: Signal attenuation  (Read 1488 times)

0 Members and 1 Guest are viewing this topic.

Offline eamoexTopic starter

  • Contributor
  • Posts: 44
  • Country: ch
Signal attenuation
« on: April 17, 2019, 08:41:23 am »
Hello All,
I'm an amateur electronic musician and I use a miniature modular synth. For those who don't know, this type of synth doesn't store particular settings for a creative sound; instead it relies on physical wire connections through a series of jacks using the many ports available on the machine itself, which mutally send and receive voltage-controlled signals, which in turn change frequencies of oscillators, or threshold of filters, etc... In other words, the control is analog, where, for instance, you have relations like octave-per-volt, or frequency-per-volt, etc...

https://cdn.korg.com/us/products/upload/7ff66deca06d37565b026ce984f059f2_pc.jpg

The thing is, the synth can receive and process voltage-controlled signals from external gear, but the standards to shape the signals are not very uniform. I have an external controller which sends signals oscillating between -10 and +10V, but the synth can only receive signals ranging between -5 and +5V.

Questions: could I easily modify a cable which would "attenuate" the signal, i.e. lower the potential from 20Vpp to 10Vpp, without losing the timing and integrity of it?

I know how to build a simple linear voltage regulator circuit, but that's for power, not signal. Not sure how to go about this one. Everything I found on the topic is not well informed and stem from people who, like me, don't really know how to do this right.

There's this: https://koma-elektronik.com/?product=koma-attenuator-cable, but it looks so sleek and simple that I can't help thinking I could build a dozen myself for a fraction of the price!

Any advice welcome!
 

Offline johnwa

  • Frequent Contributor
  • **
  • Posts: 255
  • Country: au
    • loopgain.net - a few of my projects
Re: Signal attenuation
« Reply #1 on: April 17, 2019, 08:54:28 am »
Hi,

Yes, you should be able to just use a passive voltage divider - one resistor in series with the signal lead, and one across the output of the cable. For the voltages you have given, the values are easy - just make both resistors equal in value. The only issue you will have is with impedances: for best accuracy, the resistors need to be small compared to the input impedance of your synth, but large compared to the output impedance of your external controller. Try 2x 10kOhm, and experiment from there. You can overcome this problem using a buffer amplifier, but this will add complexity.
 
The following users thanked this post: eamoex

Offline eamoexTopic starter

  • Contributor
  • Posts: 44
  • Country: ch
Re: Signal attenuation
« Reply #2 on: April 17, 2019, 11:39:45 am »
Thanks johnwa. Voltage divider, I know that. Haven't even thought about it. Plus it's real easy to test. Will do and report.
 

Offline eamoexTopic starter

  • Contributor
  • Posts: 44
  • Country: ch
Re: Signal attenuation
« Reply #3 on: April 18, 2019, 08:16:57 pm »
Okay, haven't been testing yet. But I've measured actual voltages and read the doc. Actually I need to go from 14V to 5V. I DuckDuckGoed this online calculator:

http://hyperphysics.phy-astr.gsu.edu/hbase/electric/voldiv.html

I knew about voltage dividers but this is useful to understand the influence of RL and the sum of R1+R2 on Vout under load. Now, here is the question: is there a method to easily measure RL for a given device?

EDIT: oh yes and how do I know what power values I should remain under? I see it's pretty much possible to fry something here with too low R1+R2 values.
« Last Edit: April 18, 2019, 08:19:36 pm by eamoex »
 

Offline eamoexTopic starter

  • Contributor
  • Posts: 44
  • Country: ch
Re: Signal attenuation
« Reply #4 on: April 19, 2019, 11:48:38 am »
Voltage divider with 2x10K worked great. Synth is rocking and in tune. No need to explain about impedance (which I have a hard time understanding).

Thanks for your help!
 

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 4135
  • Country: 00
Re: Signal attenuation
« Reply #5 on: April 19, 2019, 12:17:42 pm »
voltage divider doesn't keep proper impedances. With impedance mismatch you can get amplitude mismatch and even distortions, because circuit may not support low or high impedance.

To do proper attenuator you're need to know input/output impedance.

You can calculate proper attenuator here for example:
https://www.microwaves101.com/calculators/858-attenuator-calculator

In your case (10 Vpp to 5 Vpp) you're need 6.02 dB attenuator (att = 20*log10(5 / 10)).

For example, if your circuit has 600 Ohm impedance, you can use tee-attenuator:


with
R1 = 199.982 [Ohm]
R2 = 800.092 [Ohm]

you can round it with 200 and 800 Ohm resistors.

If your system has 50 Ohm impedance, you can buy 6 dB attenuator here:
https://www.aliexpress.com/item/2W-SMA-DC-6GHz-Coaxial-Fixed-Attenuators-Frequency-6GHz-SMA-Fixed-Connectors/32956090844.html

Note: input side is marked with "SU" mark. You can apply max 2W power to this side.
« Last Edit: April 19, 2019, 12:30:48 pm by radiolistener »
 
The following users thanked this post: eamoex

Offline Doctorandus_P

  • Super Contributor
  • ***
  • Posts: 4004
  • Country: nl
Re: Signal attenuation
« Reply #6 on: April 19, 2019, 10:54:50 pm »
Halving the amplitude is 6dB, which is a bit cumbersome for a passive divider for audio.
Controlled Tee Attenuators simply do not work with audio because audio devices do not have standarized or controlled input and output impedances. Audio is also low frequency, with relatively short cables and transmission line effects do not apply.

I have made some plugs with a male and female cinch connector and 2 resistors, and they work reasonably well.
For a 6dB divider I would use 2 560 Ohm resistors.
Most audio sources are capable of dirving loads as low as 600 Ohm, and with 2 560 Ohm resistors.
For more attenuation make the first resistor higher, and the 2nd lower. This will get the final attenuation closer to the theoretical values for "typical" audio impedances.

Another thing to remember is to put the attenuator on the end of the cable. ( on the receiving side).
If you put the resistors on the "sending" side of the cable the resistors form an RC filter with the cable capacitance and high frequency signals (for audio) may get more attenuation than low frequency signals.
There are still 3 orders of magnitude between 20Hz and 20kHz.

 
The following users thanked this post: cjs

Offline radiolistener

  • Super Contributor
  • ***
  • Posts: 4135
  • Country: 00
Re: Signal attenuation
« Reply #7 on: April 20, 2019, 04:20:00 pm »
Audio is also low frequency, with relatively short cables and transmission line effects do not apply.

yes, for short wires and low frequency, unmatched transmission line effect will be almost not visible. But low impedance load may change amplitude on the output driver and may change it's mode which may leads to distortions.

There is audio standard 600 Ohm impedance. But nowadays electronics often doesn't comply with 600 Ohm in order simplify and make schematic cheaper. They often use low impedance output (60 Ohm for example) to avoid significant amplitude change with connected load. So, there is need to check what output and input impedance actually is.
 

Offline calli

  • Regular Contributor
  • *
  • Posts: 100
  • Country: de
  • 3D Blender and Maker
    • Blender Buch
Re: Signal attenuation
« Reply #8 on: December 01, 2019, 05:28:27 pm »
Just need to add:

http://www.doepfer.de/DIY/a100_diy.htm

Helped me a lot! For all coming here searching for modular synths.

C!
Carsten Wartmann: Make Magazin DE - Autor - Dozent - 3D - Grafik - Maker
http://blenderbuch.de/
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf