The open source Jupiter-8-4-ALL project

From circuitbending to homebrew stompboxes & synths, keep the DIY spirit alive!

Moderators: lisa, luketeaford, Kent, Joe.

KSS
Ultra Wiggler
Posts: 873
Joined: Mon Jan 25, 2016 7:28 am

Post by KSS » Fri Dec 21, 2018 9:35 am

Have not looked at Tom's code so any statement about what his code does is unconfirmed. Unless he has implemented some sort of slop strategy,and even if he has, many do fail to replicate the nonlinear and peculiar little things which analog circuitry does inherently.

In a polysynth this matters in a very large way. The first digital polys are nearly universally regarded poorly unless upgraded with modern uP's and code. Because while it is simple to digitally generate an accurate ADSR shape, it is far more difficult to model all the little errors of the analog circuits.

Was impossible in the old synths with the level of digital available then. And is now still very hard because the effects are either not understood or characterized enougfh to model well, or are not ocnisdered important enough to do so. Some gradient between these two will always exist in any digital emulation of analog and any miss may be heard in the end result.

More important in polys than monosynths too. Because the same way we enjoy a certain type of oscillator error grouping and despise a very similar but different type, the same is true for the envelopes. Perfectly matched envelopes sound poor. Poorly mismatched envelopes, intentional or not, also sound poor. There is little disagreement about this as can be seen in hundreds of posts online across many forums.

KSS
Ultra Wiggler
Posts: 873
Joined: Mon Jan 25, 2016 7:28 am

Post by KSS » Fri Dec 21, 2018 9:37 am

ACME wrote:No, I dont think so. Before I decided to design circuit boards, I built the entire VCO on a breadorad, with multiturn trimmers. The setting of the VCO was very cumbersome and took a lot of time, because you will screw and screw and screw until due to reach the right value. The fine-tuning may be easier then, but is not necessary because it goes just as well with good quality single-turn trimmers and a calm hand. The last few cents are corrected by the autotune.
Quoted for truth. So very glad to see this accurate assessment WRT multi-turn trims.

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Post by dikkietrom » Fri Dec 21, 2018 10:06 am

KSS wrote:Did see the lushprojects link is using the Falstad sim. You'll do yourself a favor to upgrade from that one to something like LTspice. Falstad is fun and can be useful ,but fails in ways that are not obvious and can be misleading even moreso than sims almost always are anyways.

And yes, the A clock gens use 150pF timing caps while the D,R clock gens use 47pF.
Ok you found it, I use it to see and show what is going on conceptually. Totally agree on the fact that it does not mean anything besides that. Because of the realtime aspect I prefer it over spice.

Thanks I did not spot the caps differences yet.
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Post by dikkietrom » Fri Dec 21, 2018 10:46 am

KSS wrote:Have not looked at Tom's code so any statement about what his code does is unconfirmed. Unless he has implemented some sort of slop strategy,and even if he has, many do fail to replicate the nonlinear and peculiar little things which analog circuitry does inherently.

In a polysynth this matters in a very large way. The first digital polys are nearly universally regarded poorly unless upgraded with modern uP's and code. Because while it is simple to digitally generate an accurate ADSR shape, it is far more difficult to model all the little errors of the analog circuits.

Was impossible in the old synths with the level of digital available then. And is now still very hard because the effects are either not understood or characterized enougfh to model well, or are not ocnisdered important enough to do so. Some gradient between these two will always exist in any digital emulation of analog and any miss may be heard in the end result.

More important in polys than monosynths too. Because the same way we enjoy a certain type of oscillator error grouping and despise a very similar but different type, the same is true for the envelopes. Perfectly matched envelopes sound poor. Poorly mismatched envelopes, intentional or not, also sound poor. There is little disagreement about this as can be seen in hundreds of posts online across many forums.
This is a very interesting subject. I think much of the digital generated stuff would be more spot on if the programmers would know more about what is actually going on. How many vst's implement descaling per voice instead of detuning per note, which would help a lot for analog feel, I don't know any.

Totally agree about the env's, that's why I want to use at least a cap instead of lookuptables and was wondering about the clock stability of the JP4 but I see tempco's so there we go, crystal not an option ;)

The question (for me) is if an R01 is the IC version of the JP4 discrete env's, like the vcf. AMsynth also think so it seems https://amsynths.co.uk/home/products/mo ... -adsr-vca/ , if so then there is no question about what has to be done. It makes perfect sense, why would they come up with a complete different module design while everything stayed basically the same module wise just two years later. With a sensitive enough scope the steps might be visible in the slope of an R01, which would give a final answer.
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

User avatar
ehafh
Common Wiggler
Posts: 240
Joined: Thu Jan 22, 2015 11:07 pm
Location: illinois, usa
Contact:

Post by ehafh » Fri Dec 21, 2018 10:52 am

This is exciting :cloud:

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Post by dikkietrom » Fri Dec 21, 2018 11:51 am

wip interface, vectorized so can be scaled to any size while staying sharp. Can be used for pannel cutting and as a software interface. Will first make it work as a software JP with analog engine.

Image



Image
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

KSS
Ultra Wiggler
Posts: 873
Joined: Mon Jan 25, 2016 7:28 am

Post by KSS » Fri Dec 21, 2018 3:11 pm

dikkietrom wrote:It makes perfect sense, why would they come up with a complete different module design while everything stayed basically the same module wise just two years later.
need to put that two years into some perspective. Why did CRTs effectively disappear in the space of two years? Why did CDs disappear in a short timeframe? Why doesn't everyone have a blackberry instead of an iphone?

The two years you're talking about were in a similar change point as all these. The development of technology is not usually linear. But rather fits and starts. two years at one point can be the same as two weeks at another not too far calendar removed from the first. Synth developers were making a huge amount of progress in a still new field during the two years you've highlighted. What was available near the end of theat time vs the beginning of that time answers the question. Simply put, because they could do things differently. With options they didn't have two years earlier. Heady times for those of us who lived through them.

KSS
Ultra Wiggler
Posts: 873
Joined: Mon Jan 25, 2016 7:28 am

Post by KSS » Fri Dec 21, 2018 3:41 pm

Re: AMSynth's JP4 EGs, You'd have to ask Rob to be sure, but I don't read his saying that the five ICs used in the JP4 became the RO1. Both EGs, yes, but the topology inside is different. For the reason outlined in my last post. Same reason a dog does what he does. :razz:

User avatar
emuarc
Common Wiggler
Posts: 181
Joined: Sat Oct 02, 2010 4:23 pm
Location: Horsham, West Sussex

Post by emuarc » Sat Dec 29, 2018 1:13 pm

Hi,
Great project and thread, keep up the excellent work!

I have some info on the JP4 envelope generator here:
AM8108
The Roland design is innovative but very heavy on component count and really needs log not linear pots for control. Having individual trimmers for each envelope stage in a polysynth is both mad and exciting! I am not seeing or hearing an attack as fast as 0.6ms more like 2ms, but if someone could measure a real one that would be cool..

The IR3R01 replaces the 5 ICs and could be a replica of the JP4 design as the IR3109 was of the BA662’s.

The JP4 VCO is also a unique design, which was a dead end but very interesting musically. The JP8 VCO is based round the 100M VCO which was an improvement over the System 100 and SH5 thanks to JFet op amps being mainstream. There is more info on the JP4 VCO here:

AM8400

I have replaced the ua726 heated transistor pair with contemporary expo generators and tempco in all my Roland clones. This gives very good tracking and allows for hf trim.

This research is leading to a set of AMSynths commercial products in 2019. Euro Rack, pcbs and complete synths. I’ll post some more envelope timings soon.
Last edited by emuarc on Fri May 10, 2019 4:02 am, edited 1 time in total.
Rob Keeble
Owner & Designer
AMSynths

Tiergrinder
Learning to Wiggle
Posts: 2
Joined: Fri Jan 25, 2019 5:10 pm

Post by Tiergrinder » Wed Jan 30, 2019 3:47 am

Just founded this thread and I'm super happy that there are new JP8 projects turning up :yay:

I'm definetly going to follow this thread and learn more about the inner works of the Jupe too 8-)

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Post by dikkietrom » Sat Apr 06, 2019 10:03 pm

Apologies for the silence, unfortunately I cannot give this the priority it deserves right now. I can only supply voice boards, bom and instructions. The bittersweet symphony took over again. To be continued.
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

User avatar
emuarc
Common Wiggler
Posts: 181
Joined: Sat Oct 02, 2010 4:23 pm
Location: Horsham, West Sussex

Post by emuarc » Thu May 09, 2019 5:47 am

Hi,
I have started researching the IR3R01 envelope chip in more detail, and there is a post on my website to record progress to a working replica here:
https://amsynths.co.uk/2019/05/09/roland-ir3r01-clone/.

If anyone has ideas or real world measurements please comment on the post or here. My initial conclusion is that the IR3R01 is closer to the Jupiter 4 ADSR than a CEM/SSM chip, and that it is oscillator based.

I am also making progress on a single voice version of the Jupiter 4 which fits into a 60HP Euro Rack case. Designs will be in Github as Open Source Hardware as well as commercial products.
https://amsynths.co.uk/2019/05/04/jupit ... ard-recap/

Image
Rob Keeble
Owner & Designer
AMSynths

User avatar
ACME
Learning to Wiggle
Posts: 35
Joined: Wed Jan 17, 2018 8:39 am
Location: Germany

IR3R01 Plug In Replacement

Post by ACME » Thu May 09, 2019 12:37 pm

I have already made a suitable analogue plugin replacement for the ir3r01 for my Jupiter8 clone. My clone has all the functions that the IR represents. For A-, D-, R-times and S-level I have provided separate settings. The general timing is set on the ModuleBoard. Env.Key Follow is also on board.

https://www.dropbox.com/s/z2ynpymxaxciw ... 3.JPG?dl=0
(Uploading images is currently not working.)

Markus

User avatar
emuarc
Common Wiggler
Posts: 181
Joined: Sat Oct 02, 2010 4:23 pm
Location: Horsham, West Sussex

Post by emuarc » Thu May 09, 2019 2:23 pm

Very cool using the 3310 as a basis. How did you use the modulation input to vary the timing? I was aiming to get the replica into a 16 pin footprint...
Rob Keeble
Owner & Designer
AMSynths

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Post by dikkietrom » Sat Jul 13, 2019 3:40 am

emuarc wrote:Hi,
I have started researching the IR3R01 envelope chip in more detail, and there is a post on my website to record progress to a working replica here:
https://amsynths.co.uk/2019/05/09/roland-ir3r01-clone/.

If anyone has ideas or real world measurements please comment on the post or here. My initial conclusion is that the IR3R01 is closer to the Jupiter 4 ADSR than a CEM/SSM chip, and that it is oscillator based.

I am also making progress on a single voice version of the Jupiter 4 which fits into a 60HP Euro Rack case. Designs will be in Github as Open Source Hardware as well as commercial products.
https://amsynths.co.uk/2019/05/04/jupit ... ard-recap/

Image
For some reason I was not notified about these replies. I have read your article. I think it makes perfect sense to replace the logic components of the JP4 design with one pic but most certainly don't use lookup or calculated values. Use an external cap to keep the vibe. The tolerances of the caps will make it work. Mind you the end game is polyphony, the subtle differences between the caps will specifically start to shine when used polyphonic I think. It would be a waste to discover that at the end. We have the same hunch regarding the JP4 design, it would be wise not to ignore that.
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Re: The open source Jupiter-8-4-ALL project

Post by dikkietrom » Sun Dec 29, 2019 4:32 am

Some heavy weights joined in. The cv controller is ready and fully functional. The voice boards are close to being assembled for the first sound tests and the enclosure is in final design stage. Coming time will be connecting dots.



The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Re: The open source Jupiter-8-4-ALL project

Post by dikkietrom » Tue Dec 31, 2019 12:19 pm

Unfortunately we did not make it to audio this year to let you hear, this is the assembly state of the first voice proto now.

Image



This is the midi-cv controller proto with which you can control 1 to more voices. The multitimbrality, is determined by the number of these control boards. The number of these is limited by the space of the box. The pots on this proto can be dynamically assigned, with which the jp8 can already be fully operated. There is also a software interface but the end goal is of course to have the original hardware control layout, possibly with keyboard, although I don't need it myself.

Image
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Re: The open source Jupiter-8-4-ALL project

Post by dikkietrom » Fri Jan 03, 2020 2:41 pm

Spend today cutting the panel and aligning the panel pcb against the cut, which is pretty tedious, got half way. The pots and switches need to be done amongst a few single items. The plan is to have the panels cut, anodized and laser printed. For single piece or small series it is the most efficient but in my opinion it looks a lot better than coating and and printing while also being more durable.

Image

Image

Found this button, it is not exact but pretty close and they have all needed colors.

Image
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Re: The open source Jupiter-8-4-ALL project

Post by dikkietrom » Wed Jan 15, 2020 2:22 pm

decided to go full blown, top is a single sheet fold. As is the board mount panel. Under the panel board is the voice distribution, it will be 4 timbres x 8 = 32 voices max. Image

Top panel is on order it will arrive the 23th, raw alu of course
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1162
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: The open source Jupiter-8-4-ALL project

Post by devinw1 » Wed Jan 15, 2020 4:23 pm

Who did end up having fab that sheet metal? That's a beast of a part.!

Any treatment for the back corners where the top meets the side, or will it just butt up?

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Re: The open source Jupiter-8-4-ALL project

Post by dikkietrom » Thu Jan 23, 2020 1:42 pm

yo eat this haha, $70 a piece, at 10 $30 each ,insane
You do not have the required permissions to view the files attached to this post.
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

dikkietrom
Common Wiggler
Posts: 160
Joined: Sun Aug 11, 2013 10:59 am

Re: The open source Jupiter-8-4-ALL project

Post by dikkietrom » Thu Jan 23, 2020 1:43 pm

devinw1 wrote:
Wed Jan 15, 2020 4:23 pm
Who did end up having fab that sheet metal? That's a beast of a part.!

Any treatment for the back corners where the top meets the side, or will it just butt up?
you made my day man, thanks :miley:
The open source Jupiter-8-4-ALL project
Desperately seeking IR3R01

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1162
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: The open source Jupiter-8-4-ALL project

Post by devinw1 » Thu Jan 23, 2020 2:32 pm

$30 for that in qty 10???. You're joking right?

What country? :woah: :woah: :woah:

User avatar
devinw1
Super Deluxe Wiggler
Posts: 1162
Joined: Tue Aug 07, 2018 11:20 am
Location: Portland, OR
Contact:

Re: The open source Jupiter-8-4-ALL project

Post by devinw1 » Thu Jan 23, 2020 4:05 pm

Nm I see the box. Very surprised they can do that so cheap in the EU. But you saved on shipping that way too.

Now to figure out how to finish it eh?

User avatar
Jackdamery
Wiggling with Experience
Posts: 283
Joined: Sat Mar 16, 2013 3:39 pm

Re: The open source Jupiter-8-4-ALL project

Post by Jackdamery » Fri Feb 14, 2020 8:15 am

Through the courtesy of

- Mr Ikutaro Kakehashi, Roland
- Grenoble Institute of Technology, http://kicad-pcb.org/
- Alfons Wirtz, freerouter.net
- Jack Damery, An in depth look at the Roland Jupiter 8's VCO...
Yooo, just saw this. Nice. Are you still in desperate need of IR3R01s?

I mapped out most of the JP8 voice boards in Eagle and had the PCBs printed. They worked well.

I got stuck on trying to make a DIP IR3109 clone. I believe I made another post about it. This was all a few years ago though.

Post Reply

Return to “Music Tech DIY”