AOUT + AIN + EURO

[latigid on]

This is a general interest/info post for a new(ish) CV concept.

The green PCBs are 4 channel AOUT and AIN boards. Either one stacks with one blue "Control" board.

 

 

AOUT: MAX525 based. I realise that @Altitude also has an AOUT in the pipeline, so I want to explain the pluses for both.

@Altitude

• 8 channel
• more compact (minimum rack real estate)
• DOUT onboard
• lots of clocks and triggers
• US shipping

 

@latigid on

• more spacious (a bit more ergonomic)
• switchable ranges (0-5, 0-10, +/-5V)
• manual fine tune control
• through hole assembly (no fine pitch SMT)
• new amplifier circuit with precision references and compensated outputs
• EU shipping
• matching panel design with AIN (see below)

Status: I have most of the parts on hand, so this should be tested within a few weeks.

 

AIN: 

• switchable ranges (0-5, 0-10, +/-5V) to optimise ADC levels
• protection against overvoltage (op amp limiting)
• attenuator -- can set to clip the op amp input for interesting CV shapes!
• normalisation of the inputs to +5V -- can use pots as manual CV sources
• 3v3 power from a Core or an onboard Vreg

Status: tested as working from an older proto. Waiting on a few parts to complete it.

 

Control:

• Eurorack compatible 
• Accepts either "Thonkiconn"  (PJ301M-12) or PJ301BM jacks
• 4 channel AOUT or AIN
• 4 channel gate/trig/clock with LED
  • LEDs are driven from a Schmitt trigger so shouldn't interfere with your power rail
• 1 clock can be set as an input with CV protection and conditioning (even works with a sine wave!)
  • this conditioned clock can be sent to a Core e.g. MBCV
• DPDT (7mm) switches to engage the range functions:
  • switch in a resistor to adjust gain settings
  • switch in an offset for bipolar operation
    • (you could have +/-2.5V or even +/-10V if desired)
• Duo LED to indicate current CV range
• configurable adjustment pots
  • fine tune with diode deadband for AOUT
  • attenuator for AIN
  • note: the 9mm Alpha pots fit, but I have some "long shaft" versions on order. They have a knurled shaft and pointer which doesn't require a knob.
• mounting for a DOUT R5 board (see below)

Status: needs to be tested with the AOUT and AIN boards, but the mechanical fits look okay.

 

^^ just thrown together, I suppose shorter standoffs would be possible.

 

My design goal for an MBCV example is 2*AOUT and 2*AIN:

• 8 CV out
• 8 gate
• 8 CV in
• 7 clock out
• 1 clock in

You'd have 200mm of PCBs, which is around 40HP of Euro space.

 

I expect boards could cost 5-6 EUR each.

 

 

 

 

 

Edited August 3, 2016 by latigid on

[Phatline]

3 hours ago, latigid on said:

• 4 channel AOUT or AIN
• 4 channel gate/trig/clock with LED
• mounting for a DOUT R5 board (see below)

hi.

what meen "OR" exactly? > do i have to use a other "Sandwitch-PCB" behind the frontpanel when using AIN or AOUT?

Do i need a Front-PCB @6€ + a AOUT-PCB@6€ to get a 4 AOUT+Trigger-Combi?

What DAC do you plan > a 4Channel? what type?

[latigid on]

1 hour ago, Phatline said:

hi.

what meen "OR" exactly? > do i have to use a other "Sandwitch-PCB" behind the frontpanel when using AIN or AOUT?

The idea is you use:

one Control PCB + one AIN 
-- OR --
one Control PCB + one AOUT
(you can't use both in the same "sandwich")

The Control is flexible in that in can be configured for either; you stuff different resistors/diodes/solder jumpers, while the SIL headers connect to different parts of the respective PCBs.

 

1 hour ago, Phatline said:

Do i need a Front-PCB @6€ + a AOUT-PCB@6€ to get a 4 AOUT+Trigger-Combi?

Yes. Plus a some sort of DOUT (e.g. SmashTV, MBCV v2).

 

1 hour ago, Phatline said:

What DAC do you plan > a 4Channel? what type?

MAX525. It is expensive but has proven itself as a good DAC. I soldered two AOUT_NG modules, one doesn't work and for the other, channel 1 is non-linear viz. my SEQ tuning. I don't know if that's a problem with the chip; maybe I overheated it whilst soldering?? But checking on Mouser, MAX525 is one of the cheapest DACs available in DIP. There are a few others, but I couldn't find many (any?) with serial out (Tx). It's still listed in active production and if there's sufficient interest maybe I can look at a bulk order. I'd also like to chat with TK. about a possible MUX/DEMUX board. Thus you might have 32 (64?) inputs for ADC and 32 (64) outputs for DAC using SMP08s, and a 74HC595 to generate chip select addresses.

Best,
 

[Psykhaze]

Multiplexiiiiiiiiiiiiing <3

[Rowan]

I like the universal IO board concept you've come up with here, well done!

Your choice of using the MAX525 is understandable as its already supported by the MBHP. I assume that you've come across mxmxmx's Ornaments+Crime module? He's using the  TI DAC8565 which has 16 bit precision. I do understand the trade off here of 4 bits for THT but this is worth it in this case.

[Psykhaze]

wow, usually i'm not for smt but i must admit this dac8565 from TI might worth the revisiting,moreover looking @ the price&specs (16bits precision for 15€ @ Mouser)

Edited August 5, 2016 by Psykhaze

[latigid on]

I see two problems with the DAC8565: no DOUT for serial chaining, and TSSOP is fairly tedious to solder. Those 4 extra bits aren't too useful considering the error of.. 4 bits! 

I'm fine with designing and testing a replacement if one exists!

[latigid on]

MAX5500 (SSOP) could work.

[latigid on]

By no means is it impossible to solder SSOP, but the difference in pitch is significant:

 

 

[Rowan]

I've come to like working with these type of SMT devices. With a bit of practice I've managed to get very good results, indistinguishable from machine placed components.

My "technique" is to solder all the pins not worrying about solder bridges, the main thing being that all the pins are connected to the pads then use desoldering braid to clean up the joints. 

This is works every time for me even on TSSOP ICs. 

Edited August 5, 2016 by Rowan

[Psykhaze]

After a look @ mouser , many nice DACs like :

• AD5676BRUZ
• MAX5307
• MAX5591  

But all TSSOP unfortunately

Edited August 5, 2016 by Psykhaze

[latigid on]

AD5734

[Rowan]

Why not use something like this for the DAC?

 

 

I'm sure someone on this forum could/would offer a SMT service be it free or paid.

alternatively, you could have the PCBs produced with the DAC pre soldered. Some PCB manufactures offer this service. 

Edited August 5, 2016 by Rowan
Removed eBay link, added image

[Psykhaze]

ebay link dead

[Psykhaze]

Also @latigid on If we feel like choosing a new DAC in small package, why not trying to have 8 inputs like AD5676 or MAX5591 , as they are both unexpensive?
Bests,
JK

[Psykhaze]

@Rowan Adapter Idea is pretty, i'm almost sure some bulk serie could by done by midiboxer / Third part company for almost cheap as you say. 

By my side, a 8 outputs DAC on this kind of adapter would be THE right choice.

Edited August 5, 2016 by Psykhaze
mistype

[Rowan]

The adapter PCBs themselves are extremely cheap, the ones I pictured are about £2.95 for 20!

To be honest, I would happily solder a few dozen ICs to these adapters if a bulk buy of the DACs was arranged. They could be sold with the AOUT PCB as a package. Let's face it, if you are going to by the AOUT board you're going to have to by the DAC anyway.

The PCB could be redesigned so that a SMT DAC can be soldered directly to the PCB but also has holes to accommodate such an adapter.

[Psykhaze]

DIP/PDIP DACs are almost more expensive and hard to get anyway.

As say Rowan, better face the constraints we will have to face anyway, eg choosing a pretty good unexpensive replacement for MAX525.
And it seems that the right solution to this issue goes through SSOP/TSSOP packages unfortunately.
The idea of the adapter sounds very good to me, because it would allow to use DAC in the traditionals handy DIY-friendly DIP sockets. 

Better organize ourselves once the choice is fixed on the DAC to have some presoldered versions on adapters.
 

[latigid on]

(T)SSOP plus an adaptor is a good idea. I'm not sure if pre-soldered products may be sold in Europe(?). I would prefer to keep the chip coupled as closely as possible to the PCB (+1 on the dual use pads), but it looks like time is up on DIP DACs as you say.

Quad vs. Octal: I'd like to keep the quad layout if possible, although it could be doable to have an octal DAC and two quad control panels. If I can avoid re-routing the DAC board too much I'd be happy :).

Hopefully you can read that.

MAX5500A:

• Quad 12-bit
• ±0.25 (worst case ±0.75) LSB integral non-linearity 
• ±1 LSB differential non-linearity
• 3.5mV offset error
• 6 ppm/C offset-Error Tempco
• Typical -0.3 LSB gain error
• 1 ppm/C gain-Error Tempco
• Similar pinout to MAX525

-- in fact looks basically like the same die as MAX525

 

AD5676B:

• Octal 16-bit
• ±1.8 (±3 worst case) LSB integral non-linearity 
• ±0.7 LSB differential non-linearity
• ?0.75 (worst case ±2) offset error
• some other parameters are given as percentages
• cost is the same as 2* 5500
• chip layout is a clockwise arrangement: a bit annoying to keep the same order (A B C D || H G F E)

not impossible.

 

MAX5590B

• 'A' variant is not recommended for new designs... not a good sign
• Octal 12-bit
• ±2 (worst case ±4) LSB integral non-linearity 
• ±1 LSB differential non-linearity
• 5/25mV offset error
• 5 ppm/C offset-Error Tempco
• ±20/±40 LSB gain error(? 2/4?)
• 1 ppm/C gain-Error Tempco
• chip layout is a clockwise arrangement: a bit annoying to keep the same order (A B C D || H G F E)

 

AD5724RBREZ

• Quad 12-bit
• ±1 LSB integral non-linearity 
• ±1 LSB differential non-linearity
• ±6mV offset error
• 4 ppm/C offset-Error Tempco
• 4 ppm/C gain-Error Tempco
• actually a better pin arrangement
• expensive

 

Once the parts arrive I will need to test the MAX525 AOUT to see if the scaling circuit works well in general. Then we could consider an AOUT_3G (?) module.

 

 

[Psykhaze]

MAX5500A / AD5676B sounds clearly the 2 best choices.
By my side i'd like to try the AD5676 to see what an octal 16 bits could give, for the experimenting fun.

I like the name "AOUT_3G" . After some breadbord tests , maybe having 2 tests variants could be an idea (dual Quad 12 bits / octal 16 bits).
@latigid on :to me, MAX5500A sounds the best for your case, eg adapting to an existing routing.

[Rowan]

The form factor of the DAC really shouldn't be a consideration in your design, the main concern should be accuracy. Choose a chip that offers the best accuracy then design around it.

Number 8 wire theory at its best (@latigid on, I'm also a kiwi by birth;)

[latigid on]

Sure thing, hence "not impossible."

Everyone can go ahead and design their own DACs, but if there's something similar or even equivalent to one of the existing DACs, integration into MIOS should be easier. E.g. if MAX5500 follows the same code protocol as MAX525 (I need to check), then nothing has to change and it's hopefully no extra work for TK. If the pinout is the same, then I just need to get a chip + adaptor and test it in the current PCB (or Altitude could try with his AOUT boards already prepared for SSOP MAX525).

[Psykhaze]

1 hour ago, latigid on said:

Everyone can go ahead and design their own DACs, but if there's something similar or even equivalent to one of the existing DACs, integration into MIOS should be easier. E.g. if MAX5500 follows the same code protocol as MAX525 (I need to check), then nothing has to change and it's hopefully no extra work for TK.

TK is not the only one who can code =) i do not intend to re create the wheel but as rowan said, Accuracy should be the main design rule. That explains my interest in a 16 bits octal DAC, moreover that i did not started PCB work yet

1 hour ago, latigid on said:

If the pinout is the same, then I just need to get a chip + adaptor and test it in the current PCB (or Altitude could try with his AOUT boards already prepared for SSOP MAX525).

If you want to stay electronic side,as i stated, MAX5500 should be the best to adapt to exisiting quad outs max525 designs (12 bits res,same outputs number,pinout similar...).
Maybe just an adapter PCB to design to replace max525 in exisiting PCB should be enough to adapt pinout?

[latigid on]

Initial doku for AIN_4 is online:

http://wiki.midibox.org/doku.php?id=ain_4

 

[latigid on]

 

Both AIN_4 and AOUT_4 are working nicely :). The range switches are especially useful for AIN, but I can see the benefit of alternative AOUT ranges with improved accuracy when used as a mod source (rather than a pitch CV). It seems a good idea to use the bipolar option for pitch, otherwise you'll always be transposing. E.g. I think C0 in MIDIbox land is set to 2.00V?