The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Synths for sale  Search Gearslutz Go Advanced
The Mopho's pulse wave...
Old 8th September 2013
  #1
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
The Mopho's pulse wave...

I've got a Mopho keyboard and it's a great little thing. But one thing puzzles me. When listening to any patch using a pulse wave, I notice the overall sound somewhat dirty sounding. It's not very nice really. All the other waveforms sound just fine.

So could anyone do me a favour and try something for me? Listen to one DCO and set the waveshape to pulse. Do you get a proper squarewave at a pulse width value of 50, or do you need to make it 49? And secondly, take the pulse width to 99, has it gone quiet or does it sound horribly grainy? On mine it sounds grainy and the even the square wave has this grainy background noise on it.

I checked inside and the CV that controls both DCOs has a random digital noise superimposed on the correct signal. Which is probably not right, but I wanted to find out whether it was just my unit or they all do it.

Many thanks,

Tony
Old 10th September 2013
  #2
Gear interested
 

I also think the Mopho pulse wave can be very dirty or 'weedy' sounding. Reminds me of the SID chip sometimes.

Using only one oscillator and setting it to value 50 it becomes a square wave, but it's also sort of dirty...there's a faint buzz in the high end for me.

At value 99 it's very thin and scrapy sounding, hasn't gone completely quiet, no. But when I modulate the pulsewidth with a LFO it does kinda go all quiet for a bit when it reaches that high spot.
Old 12th September 2013
  #3
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Thanks for that. It seems my Mopho isn't the only one.

I've got the guys at DSI looking into this for me too. And I've had a look under the hood to see what is actually causing it. I think I've found it too. It's not what I thought it would be since it seems, that on my Mokey, it's related to the internal control voltages and not the analogue voicing at all. There may well be a fix possible.

More news as I get it.

Tony
Old 12th September 2013
  #4
This might be an obvious point but have you set slop to off, and fine tuned both oscs to 0?

Then did you run the system calibration routine? I think it's hold global and press LFO 3 but please look it up.
Old 12th September 2013
  #5
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Yes, I've done the calibrate thing. Turned off all any modulation. The graininess is still there.

I've just send an audio file of mine to DSI and it will be interesting to see whether mine is just playing up or that there is something inherently wrong with the thing.

You can hear the file here:

http://www.oakleysound.com/Mopho Pulse.wav

I have traced the problem to the control voltage lines that control the pulse width on the voice chip. There is a randomisation occurring on those particular CV lines that isn't there on the other CVs. I have asked DSI to look into it because what I've found so far doesn't indicate a fault in the hardware as such.

Tony
Old 12th September 2013
  #6
Gear nut
 

I think i get this on my Mopho Keys, i will confirm tonight, i just put it down to being analog
Old 13th September 2013
  #7
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
I have found the solution.

It is not only the pulse width control voltages that have this unwanted randomisation. All of the CVs do, but it is most noticeable on the pulse widths. What got me onto the right track was the fact that all the CVs had worse noise when they were at high values. This points the finger towards a noisy reference voltage to the system's DAC.

The DAC used on the Mopho Keyboard appears to use a 5V reference derived from the +12V supply. Now, the 12V supply gets tainted by noise from the other parts of the electronics of the Mopho - which is normal behaviour of any electronic system including synths.

The problem is that the Mopho creates the 5V reference by simply dividing the +12V rail with a pair of resistors (R67 and R71). This divided down signal, complete with a portion of the noise from the +12V rail, gets sent to a voltage follower (U18b) and then sent to the DAC. Oh dear... because this noise now ends up on the CVs. The higher the CV the more noise gets through.

Solution: Filter the +5V reference voltage before it goes to the voltage follower. I did it with a decent 10uF electrolytic or tantalum cap soldered directly across R71. The reference voltage is now nice and quiet. The Mopho behaves better, not only with respect to the pulse wave, but also seems smoother in general tone too. An excellent and easy modification to do.

Tony
Old 13th September 2013
  #8
Lives for gear
 

ya cause the engineers couldn't think to clean the power feeding into a dac?

"mind blown"
Old 13th September 2013
  #9
Gear Addict
 
steffensen's Avatar
 

Quote:
Originally Posted by Synthbuilder View Post
I have found the solution.

It is not only the pulse width control voltages that have this unwanted randomisation. All of the CVs do, but it is most noticeable on the pulse widths. What got me onto the right track was the fact that all the CVs had worse noise when they were at high values. This points the finger towards a noisy reference voltage to the system's DAC.

The DAC used on the Mopho Keyboard appears to use a 5V reference derived from the +12V supply. Now, the 12V supply gets tainted by noise from the other parts of the electronics of the Mopho - which is normal behaviour of any electronic system including synths.

The problem is that the Mopho creates the 5V reference by simply dividing the +12V rail with a pair of resistors (R67 and R71). This divided down signal, complete with a portion of the noise from the +12V rail, gets sent to a voltage follower (U18b) and then sent to the DAC. Oh dear... because this noise now ends up on the CVs. The higher the CV the more noise gets through.

Solution: Filter the +5V reference voltage before it goes to the voltage follower. I did it with a decent 10uF electrolytic or tantalum cap soldered directly across R71. The reference voltage is now nice and quiet. The Mopho behaves better, not only with respect to the pulse wave, but also seems smoother in general tone too. An excellent and easy modification to do.

Tony
Wow, thanks Tony. This have been bugging me for years.
Old 14th September 2013
  #10
Lives for gear
 

I wonder if some of the other DSI synths could benefit from this kind of mod as well?
Old 14th September 2013
  #11
Gear Maniac
 

Can you post pictures about this? Deserves a sticky and applause, this is real stuff here.
Old 14th September 2013
  #12
m.o
Lives for gear
 
m.o's Avatar
 

Did it matter at all if you had the Mopho connected to USB?
Old 14th September 2013
  #13
Lives for gear
Wow just knowing this now, I will never consider buying a Mopho Keyboard. I do not have the skills to do that kind of mod.
Old 14th September 2013
  #14
Oli
Lives for gear
 
Oli's Avatar
 

Wow, that's really shoddy design. Fundamental stuff. It seems that often when I look at synth/audio schematics, there is a lot of design practice which appears really unadvisable. Makes me wonder how much of it is deliberate (for reason I haven't thought of), unaware, or just cost cutting. Something like that would cause my comms devices to fail completely.

It will be interesting to hear DSI's response.
Old 14th September 2013
  #15
Lives for gear
 

Old 16th September 2013
  #16
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Quote:
Originally Posted by m.o View Post
Did it matter at all if you had the Mopho connected to USB?
No, this is a fundamental problem caused by a poorly filtered internal voltage reference. It happens whether USB is used or not, whether midi is used or not, or even whether the supplied power pack is used.

Every internal control voltage in the synth is affected by the noise superimposed onto the DAC's reference voltage. Filter cut-off, resonance, output volume via the VCA, everything. The pulse width does show the fault more simply because that particular part of the electronics is more susceptible to this type of noise.

Tony
Old 16th September 2013
  #17
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Quote:
Originally Posted by Oli View Post
It will be interesting to hear DSI's response.
Not heard back yet.

The support person was however very friendly and helpful. I couldn't ask for more from her.

I did ask for the schematics for the board so I could look into a few other things that were bothering me. It's not that hard to reverse engineer the bits I need but it would have been easier with the circuit diagram in front of me. But the support lady told me that it wasn't company policy to give out schematics. Which I can understand.

What is interesting is that I have had a look at some online pictures of the internals of the Mopho desktop and that seems to be built the same way. Which means the modification will work wonders on the desktop version too. See my forthcoming post about my ideas about this.

I don't know about the X4 but if it does have that random grainyness on the pulse width it may well have the same 'feature'.

Tony
Old 16th September 2013
  #18
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Right for those of you that would like to do this modification, here's a bit more information:

The noise we are trying to filter out is mostly coming from the -5V charge pump power supply. The current spikes at 80kHz or so are polluting the supply voltage and the +12V low drop out regulator can't keep it clean. It is possible to filter the +12V feed to the +5V regulator with a RC network of 10R and 220uF. This will remove most of the noise on the +12V line. However, this is a more tricky modification and it still doesn't remove all the noise on the +12V that then feeds into the DAC. So filtering the DAC reference voltage is the best idea.

The Mopho Keyboard is primarily built from surface mount components using lead free solder. This makes it slightly more awkward to work on than older through hole designs. However, it is quite possible to do a nice job if you take care.

You'll need one 10uF, 16V tantalum capacitor. You can use a decent quality electrolytic if you wish, and in that case, a 10uF, 25V will suffice.

Solder the cap across the resistor R71. With the keyboard to the front of the unit, the positive terminal of the cap should be facing left. That is the negative end should go to the ground plane which is attached to the end of R71 that is nearest to the number 1 of its name.



And a closer view:



Don't pull too hard on the cap once it is soldered in place because you might pull off the solder pads of R71.

For the Mopho Desktop:

This seems to be built in much the same way with the DAC reference being taken off the +12V line. I am guessing it too will have the same unwanted noise on the DAC reference. However, the board is different in several other respects and the parts numbering is different.

Now I haven't tried this but from online photographs I can assume that you will need to solder the 10uF cap across R52. This is situated, again above the DAC, near U12. The negative pin of the cap should be soldered towards the front of the unit, ie. the positive pin will be mounted nearest the R52 legending.

Tony
Old 16th September 2013
  #19
Oli
Lives for gear
 
Oli's Avatar
 

Thanks for the great updates. (btw, nice close up photos too)

Would it be possible to put an inductor on the supply to the charge pump, to reduce noise coming back from it to the supply rails? I've done this on a charge pump before, which worked well enough to solve issues in that particular design.

As you're very experienced with synth design, building and also servicing, do you see a lot of 'interesting features' like this? I don't have much experience in audio electronics, but often see aspects of designs which appear strange to me. In this case, it not only looks wrong, but also looks like the kind of thing engineers would normally test for in prototyping.
Old 16th September 2013
  #20
Gear Addict
 
Bobby Vosene's Avatar
 

Brilliant stuff! I might try this one day - but I'm going to have to practice my soldering first...
Old 16th September 2013
  #21
Lives for gear
 

That is crazy. I really hope it was a design decision. I'd like to hear the improved version too.*

I wonder if the evolver and prophet 08 have the same thing going on.

ETA:
*and both version with some other modulations going on, especially very fast ones. I also don't recall getting this noise on my evolver.
Old 17th September 2013
  #22
Lives for gear
 
Stevism's Avatar
Awesome! It would be cool if you posted demos of stock presets, and then a user posted the same presets playing the same pattern.

really really curious to hear an A/B comparison
Old 17th September 2013
  #23
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Quote:
Originally Posted by Oli View Post
Would it be possible to put an inductor on the supply to the charge pump, to reduce noise coming back from it to the supply rails?
Very possibly. I tend to avoid using inductors simply because of their price and bulkiness. But I'm not that familiar with this style of one chip charge pump so I'm unsure what affect driving one through an inductor would have. I'd need to read up more about them.

The 10R 3W resistor and 220uF cap works very well. But although this does remove the unwanted perturbations on the +12V supply it doesn't work as well as filtering the DAC reference voltage directly. A combination of the two would be best as, right now, the 80kHz noise is having to be worked out of the audio path by the op-amps' PSRR and feedback mechanisms. I'll be looking into this at some point in the future when I get some free time.

The Mopho generates the +5V digital line from the +12V regulated supply. Personally, I would have driven the +5V regulator from the 13V wallwart directly with some additional RC filtering. That would have kept the +12V line clearer of any of the +5V noise.

Quote:
Originally Posted by Oli View Post
... do you see a lot of 'interesting features' like this?
Let's put it this way; I have never owned a synth, except the Virus KC, that I haven't modified in some way.

Tony
Old 17th September 2013
  #24
Lives for gear
 
Synthbuilder's Avatar
 

Thread Starter
Quote:
Originally Posted by Stevism View Post
... really really curious to hear an A/B comparison
I'm afraid I don't have the time to do it this week. If I get chance I'll pop some up... but really, with regard to the pulse wave, it just behaves as it should do now. The difference in the other parameters is less obvious.

Tony
Old 17th September 2013
  #25
Lives for gear
 
GeorgeHayduke's Avatar
 

Quote:
Originally Posted by Synthbuilder View Post
Right for those of you that would like to do this modification, here's a bit more information:..
Great stuff
Old 23rd September 2013
  #26
Gear Maniac
 

Does the x4 have this issue as well? I've been considering one for a while, but that audio example of the pulse wave does not sound promising...
Old 23rd September 2013
  #27
Lives for gear
 
divemedia's Avatar
 

i wonder if this is also the case in the tempest. i'll have to compare my pulse wave when i get home unless someone can do it right now.
Old 23rd September 2013
  #28
Gear Guru
 
zerocrossing's Avatar
Wow, thanks for this. It's funny, I know the sound you're talking about but it was something I heard once and then later couldn't replicate and I forgot about it... maybe thinking it was just the MoPho's character. I guess it still is, but like crooked teeth can be fixed by some orthodontic, this can be fixed with a tiny electronic component. Maybe my power is pretty clean? I'm not all that coordinated so taking this one might be a bit of a PITA, but I do have a friend who's quite skilled at electronics design who could help but I'd have to send it to him and wait until he could get to it... which with him could be a very long time.
Old 30th September 2013
  #29
Here for the gear
 

The DSI Tetra suffers from an issue, when you have low enveloped attack and decay times. A metallic spring type of noise is added to signal after the filter.

I'm wondering if this is due to the noise described in this post?

The CEM 3396 voice chips inside the Obertheim Matrix 6 (similar to the DSI chips) did not exhibit this behaviour.

Edit:
I sent this clip to the DSI Support team some time ago to demonstrate the problem:

http://www.justinvaler.com/www/tetra...normalised.wav
Old 30th September 2013
  #30
extremely interested. Desperate to hear some a/b of basic patches as mentioned earlier.
New Reply Submit Thread to Facebook Facebook  Submit Thread to Twitter Twitter  Submit Thread to LinkedIn LinkedIn  Submit Thread to Google+ Google+  Submit Thread to Reddit Reddit 
 
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump