Synthesizer Build part-65: YAMAHA CS FILTER w IG00156

 A LP, BP and HP filter in one chip. Not a filter you can just decide to build on a whim though. It uses Yamaha's own IG00156 VCF chip which is very rare and very expensive if you can find one. Like three figures expensive. This projects deals with a Eurorack version of this filter but of course it will work equally well for Kosmo sized synths.

A good friend of this website, a fellow Dutchman who happens to be an amazing psy-trance producer by the name of Jake Jakaan, signed to a top record label, who uses modules from this website gave me one of these chips. He managed to get hold of a few of them. He states this filter is great for filtering FM sounds from the TH VCO555. It's very low-mid heavy.

It was not easy to find a good schematic for this filter. In fact, all I had was the service manual for the CS-5 and a stripboard layout that someone put together which looked very dodgy and had some mistakes in it (although it did seem to work for my friend but I didn't use it).

I made a completely new schematic for this filter using the original circuit from the service manual and from that I made a new layout, small enough to fit a Eurorack system. The layout turned out to work faultlessly straightaway for which I was very grateful because the IG00156 is not a chip I'd like to blow up. The chip is actually quite robust, I found.

There are two things that are unique to this filter; it has a frequency dependent Q (resonance peak) and a gentle single pole lowpass effect. Resonance (Q) is achieved by damping rather than using a positive feedback loop and because damping will not go to zero the filter can not self oscillate.

This filter is used in all of the Yamaha CS range of synthesizers. Even in their flagship synthesizer, the CS-80. It's a two pole 12dB/Oct. state variable filter. In the CS-80 one chip is used for a lowpass filter and a second one for a highpass filter and in other CS synths like the CS-5, one chip is used for lowpass, bandpass and highpass. The filter I present here has all three functions under a switch although you could have each output go to its own socket but then you have to redesign the output stage or simply bypass the output opamp which I wouldn't advise. You should have opamps with a little bit of gain on the outputs.

Here is the schematic I made and on which I based the layouts below:

Here's a block diagram of the inside of the IG00156 chip:

Source: Yamaha IC Guidebook.

There's an in depth analysis of this filter on the Modwiggler forum: CLICK HERE to read that.

I added an audio level control on the input because this filter is rather sensitive to high volume levels. I also added a gain potmeter on the output that gives you the option to set the gain from 2 to 4 times with a 100K potmeter. That is more than enough, but if you want more gain, put in a 500K potmeter which will give you 12 times gain, or 1M which will provide 23 times gain. It will just clip. There's no use in doing that. I really wouldn't advise it.

The schematic says to use +/-15V but I tested it on a dual 12V powersupply and it works fine. The chip inputs for the cutoff and resonance functions are very sensitive and the complete cutoff range is controlled by a voltage that goes from 0 to 0,25V. Only 250 milliVolt for the full range. This is achieved by the voltage divider consisting of the 22K resistor in the cutoff line and the 470 Ω resistor between pins 7 and 8 (or pin 7 and ground really). The potmeters for Cut-off and Resonance need to be fed with +10 volts so I added 1K8 resistors to the pins where the power comes in to cut off roughly 2 volts. You can actually use other values for these potmeters because they're only used as voltage dividers but then you will have to re-calculate the values of those two resistors to make sure the pots receive +10 volts. For instance, if you use 100K potmeters you're need to use two 18K resistors.

I used this schematic to make my stripboard layout which wasn't that difficult because it's quite a simple filter. There aren't many components in it. It worked straight away although at first I could get it working because I had not wired everything up yet. I thought I had connected all the knobs I needed for testing but I forgot the V/Oct input. That needs to be connected to ground if it is not in use and once I had done that the filter sprang to life. 

VOLT per OCTAVE INPUT:

At first I had the V/Oct socket grounded through the socket switch but then I realized that isn't needed because the V/Oct input is always connected to ground via the 470Ω resistor. 

I applied a voltage divider to the V/Oct input consisting of an 18K and a 470Ω resistor, I went with an octave range of 8 octaves, meaning that the input would get 8 Volts at maximum which would need to be reduced to the same range as the Cutoff input because they all enter the same summer inside the chip (see diagram above). 18K with a 470Ω would give 0 to 203milliVolt which works out perfectly.

Because the filter can not self-oscillate anyway, the filter can not be used as a sinewave oscillator with the resonance fully open. So accurate volt per octave tracking is not an issue and therefore none of these calculations need to be super accurate, it just needs to work so that it sounds good and the filter now tracks nicely up with the octaves.

LAYOUTS:

Here are the layouts I made for this filter. As always they are verified, I used them for my build. As you can see it's a really simple project. There's only 25 resistors, a few capacitors and wirebridges and the chip sockets to put in. The biggest job will be the wiring up of the potmeters and sockets and the making of the panel.

Wiring diagram. Note that the Resonance potmeter is connected the other way around from all the other potmeters, with ground at the clockwise lug. This is usually the case in VCF's.  

Stripboard only:

It's best to use bi-polar or none polarized capacitors for the 1µF caps in the filter outputs and on the audio input. This is because we're dealing with bi-polar signals that go through the zero Volt line. The caps don't need to be this specific value. You can use anything between 1 and 10µF without problems.

As you can see all three filter outputs have a 100K resistor to ground. Together with the 1µF capacitor this forms a highpass filter with a cutoff frequency of 1.6Hz so in effect it keeps DC voltage from passing and all the other frequencies get through. Make sure you use high quality capacitors for the two 1,5nF filter caps. I used Polystyrene ones which always sound the best.

Cuts and wirebridges. As always, mark the cuts with a Sharpie or Edding pen on the component side and then stick a pin through the marked holes and mark them again on the copper side. Then cut the strips at the marked positions with a sharp hand held 6- or 7mm dril bit. Make sure you work accurately!!

And here is the bill of materials. It won't be easy to find an IG00156 chip. They are long out of production so your best bet is websites that sell rare synthesizer components. They go on Reverb for $189,- but that is top dollar. They should go for between $70,- a $100,- 

I did not include any bypass caps or extra electrolytic caps for the voltage rails. If you want to include those you need to add them to the list (2 x 100nF and 2 x 10µF/25V). They're not on the layout or schematic either but there's room enough on the stripboard to put them in over the voltage rails.

The two trimpotmeters should really be the normal kind and not multiturn trimmer. There's really no use in having multiturn trimmers because there's no need for that kind of accuracy.

NOTE: There is a Hongkong based listing of the IG00156 chip floating around on the internet selling them for €10,- Don't fall for that, it's a scam!!

PICTURES:

Here are some pictures from the build proces:

All components soldered on:

Testing:

Drilling the panel using a copy of the panel design I made in Photoshop as a dril guide:

Waterslide design applied to the faceplate and now drying on the central heating. All the creases you see will be gone by the time this is dry.

Finished panel ready to receive the pots and sockets etc. After the waterslide design has dried we cut out the holes with a very sharp hobby knife and then apply two more layers of clear acryllic lacquer and let it dry overnight.

Finished module. The module is 14hp wide (7 CM) and 3.7 CM deep.

I only had a 4 way rotary switch, that's why the HP mode on the panel has 2 settings. Lugs 3 and 4 of the switch were connected together.

Side view:

Side and back view:

Here's an oscilloscope screenshot of a squarewave wave in LP mode with full resonance applied:

CALIBRATING THE FILTER:

There are two trimpots on this filter that need to be set.

The first one is the 100K trimpot. You use this to set the throw of the Cut-Off potmeter. Set it in such a way that you get the most resolution of the Cut-Off potmeter.

The other one, the 200K, is used to set the Resonance to maximum. Turn up resonance and turn the trimmer until you're at max resonance. This will probably be around the middle of the trimmer at about a 100K.

It's best to have normal trimpots not multiturn ones. There's no need for those and the normal trimpots are easier to use.

VIDEO DEMO:

Here's the first test I did after finishing the project. This filter literally makes the room shake. If you turn up the volume (with good speakers or headphones) you will hear stuff starting to rattle in the background. It's a very bass heavy filter which really sounds great!

Here's a short demo I made with the X4046 VCO hard synced by the 555 VCO going through the filter.

Here's an other video I found on YouTube dealing with the CS-5 Lowpass filter:

https://www.youtube.com/watch?v=5F8_RfVzO2A

Here's a Facebook video of my friend Jake Jakaan using 3 of these filters in bandpass mode to create a formant filter that makes sounds akin to human speech.

https://www.facebook.com/share/v/1A2Q6Ws1AE/

Okay that's it for this article. Not a filter anyone can build alas but this website is an archive of all the modules I built myself so it certainly belongs here. I also noticed that Yamaha filter schematics that specifically deal with the CS filter are almost non existent on the internet except for the service manuals. So I hope this article will provide at least one good schematic for those looking for it.

If you have any questions or remarks about this project then please put them in the comments below or post them on the special Facebook group for this website.

If you like this content and would like to support the projects and upkeep of the website you can buy me a coffee. There's a button for that underneath the menu if you're on a PC or Mac. Otherwise you can use this PayPall Me link and cut out the middle man. All donations go straight back into new projects.

Thank you!