Need help analysing an old 8bit MDAC circuit

Help Support GroupDIY:

krabbencutter

Well-known member
Joined
Jul 11, 2018
Messages
104
Location
Germany
Hey everyone,

I've been dedicating a lot of time and effort to understand and preserve as much information about my Ursa Major Stargate 323 reverb, as possible. I've figured out most of the circuit logic but I simply can't wrap my head around the DAC configuration and would really appreciate it, if someone could help me out. I've attached the relevant bits of the schematic and here's where I'm currently at:
The DACs are 8bit current output DACs. U33, U27 & U20 produce a signed 16bit output signal, whereas U40 acts as a digitally controlled gain reference. The reference current for U40 is set at pin 14 and should be 4.01mA (+15V across a resistance of 3,74K). According to p.14 of the datasheet the full scale output current would be calculated as (X/256)*4.01mA where X is the 8bit digital input data of U40. So the maximum output current of U40 should be 3.99mA. The overall DAC arrangement seems to be the same, as described on p.9 of this application note collection (Fig.30). And as far as I understand it, the full scale output current of U33, U27 & U20 would be calculated as (iO1*R2)/R3, where iO1 equals the full scale output current of U40, R2 equals R146 and R3 equals R144/R142/R140. Since all those resistors in the schematic are 200 Ohms, the full scale output current of U33, U27 & U20 should simply be 3.99mA as well?
The signal then passes through an analog switch and into the current-to-voltage converter built around U32. The signal will either be routed to the positive or negative input of U32, depending on the sign bit. The signal then passes through another analog switch and through one of the integrators built around U23/U24. Last but not least the integrators for the left & right output channels are followed by an elliptic Low Pass Filter, a deemphasis filter and another gain & mixing stage.
Here's where my analysis falls flat:
when I simulate the analog circuitry (transimpedance amplifier, integrator, LPF & gain stage) in LTspice with a 4mA input current, I end up with a voltage gain of over 30dB and a signal that is way beyond the +/-15V supply lines.
I know it's probably a lot of information, but I really hope someone might be able to help me find the issue here.

P.S.: don't have access to my simulation file right now, but will attach it later.
 

Attachments

  • stargate_dac_output.png
    stargate_dac_output.png
    246.1 KB · Views: 13
Last edited:

krabbencutter

Well-known member
Joined
Jul 11, 2018
Messages
104
Location
Germany
I've attached the full resolution schematic and my LTspice simulation
 

Attachments

  • stargate_323_ltspice_2.zip
    238.8 KB · Views: 2

abbey road d enfer

Well-known member
Moderator
Joined
Jan 22, 2008
Messages
12,994
Location
Marcelland
I've attached the full resolution schematic and my LTspice simulation
I can't really read the schemo, but indeed the simulation shows square-wave output.
There could be many explanations:
  • Wrong assumption regarding input current
  • Error in the schematic
  • Error in your transcription of the schematic
I would suggest you measure the voltages to check if they make sense vs. the simulation.

You may try to get in touch with Cristopher Moore, who owned Ursa Major and designed all the products.
His new company is Seven Woods Audio in Belmont, Mass. He's a lovely guy and loves to discuss his products with users.
 

krabbencutter

Well-known member
Joined
Jul 11, 2018
Messages
104
Location
Germany
Thanks, I'm already thinking about buying some DAC08s on mouser so I can do some testing. My own hardware is way too precious and I'm way too afraid of slipping with a probe or sth 😅 The original DACs were sorted by Ursa Major, so damaging one of them could render my unit irrepairable.

As much as I'd love to get in touch with Christopher Moore, I'm afraid this isn't possible anymore. According to a blog post comment from 2018, he was suffering from Parkinson's disease and his website is not available anymore since late 2020.
 

Latest posts

Top