70's electric analog organ synth DIY thingy

so, i’ve found an electric organ (GEM Rodeo 61, it’s an Italian organ) from the late 70s with accompainment, drums and so, I also have the schematics
I’ve decided to turn it into a polyphonic synthetizer, since i wanted to get into synthetizers and learn more about electronics. I’ve figured out how to do most of the things like triggering the drums and making a sequencer, using the internal LFO, adding pitch bend, messing with the octaves and even cv input for the oscilator
the only thing i can’t quite figure out is changing or shaping the waveform from square to sawtooth
at first i thought all i had to do was slap a sawtooth wave oscillator instead of the square wave one, but after searching online for datasheets of the ICs under the keyboard, i’ve found that they are clock dividers that divide a very high frequency into lower, audible frequencies. But they only work with square waves, and I’ve read that whatever you send in those chips, they always output square waves

Master frequency generator aka oscillator, LFO, and divider chips

i’ve searched online for circuits that change waveforms, i found very few and just one or two worked nicely,
i tried this one, and modified it a bit, it works with all frequencies but needs adjustment for almost every frequency
WhatsApp Image 2021-11-04 at 13.32.36

slightly modified circuit

what happens if I increase the frequency

i’m taking this project of mine as a way to improve my electronics skills
so, does anyone knows if there’s a way to fix this problem? Thanks

2 Likes

Yes - what you’re looking at is a Top Octave Generator that then gets divided down to make all the notes of the octaves below and so forth.

It’s a classic organ/string machine way to get polyphony.
All the dividing is done digitally - which is why everything is a square wave.

I don’t want to discourage you, but doing the same thing with other wave shapes isn’t easy. You’d most likely have to convert each of the notes separately, so you’d be looking at 61 copies of which ever circuit you decide on.

You can turn square waves into triangle waves with an op-amp integrator circuit - but that too will need tuning for each particular frequency - you’d need to carefully select the integrator cap and input resistance for each note if you wanted the waves from all the notes to have the same amplitude. Once you have equal amplitude triangle waves you could use a rectifier similar to what you’re already doing to turn the triangles into a saw tooth.

I’d be keen to see what you get out of this so please do keep us posted!

4 Likes

thanks, i might consider doing it in a future now that i know that I have to make the circuit 49 times… (the first 2 octaves have the same tones). Maybe I could start with an octave and change the affected one with a rotary switch

one more question, in the future I’ll certainly add an envelopee for the notes’, but i’ve noticed something that I don’t like


this is the internal “envelope”, there are 3 modes that affect the attack and/or the decay, at first you might think that there’s nothing wrong, right? the “envelope” is monophonic.

what i mean

how can i make it “polyphonic”? do i need to add an envelope for every note😬? and what about filtering? i saw that Sam, in the gameboy mega machine, built a polyphonic filter and envelope, should i do something like that?

It’s paraphonic which is not an uncommon economic design choice.

The options to make it paraphonic would be to either make a filter and envelope for each note (not economical IMO) or build say 6x filters and envelope and develop a voice allocator to get 6 voices of polyphony (this is relatively hard)

My suggestion would be to embrace the limitations of the instrument and look at what you can do from there.
Some “easy” suggestions:
Better paraphonic filter with cutoff and resonance controls (ms-20 & the Moritz klein diode ladder filters spring to mind)
Chorus (research the solina strings organ)

2 Likes

I’m with @JaggedNZ here. There are so many possibilities without redesigning the wheel.
The amp and keybeds are brilliant so I’d tap in there and add fx or even another discrete synth such as the nano version of the solina string synth.
Adding a simple bypass between the organ and the amp would let you add guitar pedals or other DIY filters on an add hoc basis.
Tapping the keybed with an Arduino could give you midi or even midi to Cv also.

2 Likes

I’ve built the nano Solina and looked into the code, and my recommendation is, particularly as he has a working top octave generator, to stay well away from it. The nano just does not have the chops to compete with it.

2 Likes

True, it’s not a Powerful build, and a bit shonky up top, but tapping the keybed for midi and mixing the output with the amp’s input is good fun. I did this, building Jan’s solina into a cheap electronic piano that had midi and good speakers.
There are plenty other projects that could patch in without damaging the original organ.

1 Like

I used to have 2 organs from the early 70s. One Farfisa (pretty good brand) and a CRB Elettronica (can’t find anything about it)

The CRB had 12 LC oscillators, one for each note. On the oscillator pcb’s there was also a divider IC. I’m guessing the triangle wave eas generated in the IC as well (turns out there were a lot of organ/instrument specific IC’s back then)
The 4’ ‘pipes’ were sine wave or triangle or square. The 8’ and 16’ were square wave only. The note percussion envelope never worked properly.
This was a compact organ.

The Farfisa (80 kilogram home organ) had a 13 note top octave generator ic. 13 note means it generates both the high c and the c one octave below. The lowest 12 notes were divided down for the other octaves. There was an insane amount of switching going on. Sp12t switches and also logic gates.
Every voice (clarinet, flute, oboe, violin, cello, bass guitar and another few) had their own filter board.
It was often just transistor driven RC filters, but some were quite complex.

I’m guessing if you want full polyphony with individual envelopes for each note you would have to find a way to trigger the envelope when you press the corresponding key. Most organs just used a foot pedal for volume control (manual envelope)

In short most organs work like this:
Top octave generator>divider>formant filter>summer>volume control (manual envelope)
Good luck :slight_smile:

3 Likes

hey i’m back
3 years later and i’ve decided to resume this project since i’ve finished other more important project i had


I’ve decided to change the plan a bit, i want to get the most out of this organ, so i want to:

  • build a ±12v power supply
  • add controls for the internal lfo + cv (more like pitch bend) in for the notes
  • divide the top 4 octaves audio output + gate out for each octave (better paraphonic filter yay)
  • add a control to disable one of the 3 square waves that form the voice, also divided for the 4 octaves (see the second panel in the image)
  • vcf + adsr (more on theese later)
  • more controls on the auto chord + arpeggiator part
  • more drums and big button sequencer (i’m trying to get a better kick drum and a more tunable snare)
  • mixer + audio amplifier stage
  • effects & modulators (phaser, ring mod, distorsion, chorus…)

i’ve sorted out how to do the first 4 points, some are done some are on the way, i’ve made and tested various electrical diagrams

right now i’m working on the filter + adsr (LMNC MS20 + MK ADSR)


BUT, the filter doesn’t filter, i’ve already checked every connection and found some errors, fixed them but still doesnt work, i used two bc327 instead of the two bc556, but the resonance works. maybe i got scammed on aliexpress and they didn’t send me lm13700 idk, my local electronics supplier that is expensive as gold is out of this chip so i can’t confirm, but the tl074 are legit.

What do you think guys?

1 Like

Probability of a fake, non functioning LM13700 from AE is not negligible. You could try building one of the circuits from the datasheet to test it. Tayda still carries through hole LM13700, $4 each but at least they should work. Other sources are mentioned in LM13700 End of Life.

If it’s not the chip:

1 Like

Oh i didn’t know that the lm13700 was about to get discontinued. I’ll try to come back to my trusty electronic shop in town to see if they’ll restock, otherwise i found another seller on AE that seems to sell legit lm13700 (a rating explicitly says that he tested 2 in a VCA)
https://a.aliexpress.com/_EyqZEwN
I’ll keep you updated

It has been discontinued, it’s been out of production a couple years or so now (the through hole DIP version, that is — the surface mount version is still in production). Though as mentioned in the linked topic, there are copies being produced by Asian companies.

AE is a crapshoot; I would not place much faith in the customer reviews, and even if one customer got something usable that doesn’t mean the next one will too. I get all my chips from reliable sources, and AE is not what I call reliable. (Nor eBay, nor Amazon Marketplace. Whereas Tayda, Mouser, Digi-Key, Cabintech Global, et cetera are.)

1 Like

You can find more info, experiences, and suggestions about the discontinued through-hole and fake (or out-of-spec) LM13700 in our discussion here. In short your/our options are:

  1. Buy the cheap SMD version of LM13700 and learn to solder SMD into an adaptor. The IC and adapter come to a total of about EUR 1 together (around where I live at least) so it makes sense to learn the SMD soldering skill and buy in bulk if you’re going to use many of them in your modules.
  2. Look for new old stock from reputable sellers. These can be either large online vendors (as @analogoutput suggested) or small local shops (as I have been lucky to have around). In this case, you accept paying a premium price (EUR 3.5 to 5 in my case) and the fact that the shops will run out of stock eventually.
  3. Take a gamble and buy cheap from ebay/AE but test every IC before you put them in the circuit as fake/faulty ICs can cause serious damage. In the linked discussion above you can find instructions on how to build a simple test circuit on breadboard. I have also made a stripboard layout in case you can’t be bothered to set up the circuit on a breadboard (I will post it in the forum once I verify it…)
  4. Buy (in bulk) the Chinese through-hole version called XD13700 from AE. You might want to test those too. You are probably best testing every IC while you are at.

I have opted for 2 and 3 in the past, but 1 and 4 might make more sense if you’re going to use the IC a lot in your builds.

Buy (in bulk) the Chinese through-hole version called XD13700 from AE.

No reason to buy anything from AE. Xinluda XD13700 and HGSemi LM13700N are both available for cheap on LCSC, which is a proper supplier. (Same company as JLCPCB)

I’ve collected links here:
https://wiki.synthdiy.com/parts/otas/

2 Likes

Mouser and, to a certain extent Tayda, are not very affordable for people in Europe as they charge a lot for shipping and, in addition, you will be hit by VAT (19-25%) + processing charges (~EUR3) at the customs, bumping the total cost considerably, we’ve discussed this before.

For some reason, very small parcels from China get past the customs without charges (I think they might be bulk processing them in a transit country, Belgium or Hungary typically from my experience), so this makes ebay and AE very affordable for small orders, but whether anyone wants to risk buying ICs from there is another question…

LCSC looks indeed a more affordable option (thanks for pointing that out!). For 10 ICs I get around ~EUR4 in merchandise total and ~EUR11 in processing and postage (notice the cost…). If LCSC gets cleared in customs, that comes to EUR1.50 per unit. If not, then add another ~EUR4, to EUR1.90 per unit.

Personally, however, if I ever wanted SMD ICs I would not consider LCSC. One of the physical shops that I use can make an LM13700N order for me at their wholesale vendor (TME) and get it to me faster and cheaper than LCSC.

And then there’s option 5.: Buy 25 of them four years ago at half the current price and still have about 15 in your stash…

2 Likes

We need a flux capacitor for that though!

1 Like

And those are out of production…

1 Like

i actually wanted to start this project in 2020-2021 -_-

1 Like