 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
216.196.215.99
In Reply to: Common Grid IV Conversion - Circut Help Needed posted by kblsnbts on July 11, 2006 at 05:28:33:
Gang,There are basically 2 types of current dacs. The ones with +/- supplies where the iout = 0v in steady state or the single supply dacs (i.e. TDA1543, 1545) that have the iout = some voltage depending on VCC and their internal current sources.
I have been looking at the +/- ones because some of the better 24 bit dacs use these. I looked at this schematic and don't really see much difference between that one and just connecting a 100 ohm to ground then adding gain & output. The battery is used to bias the 100 ohms negative to the dac which should put the cathode at like 0v. The 100 ohms would also bias the tube depending on the plate voltage and current. The change in current both throught the 100 would change the bias and therefore the signal at the plate. Problem is these dacs cannot drive the 100 ohm resistor. Most of the output impedances are like 500 to 1000 ohms. So driving the 100 would be an issue.
Remember these dacs where created to drive the - input of an opamp. Which by itself is a high impedance input. BUT when adding the feedback resistor from the output of the opamp back to the - input (i.e the output looks like ground), thereby setting the imput impedance the dac is seeing is basically the feedback resistor to ground.
BUT... remember this is an active part. So preserving that can help with dynamics and such.
My ideas have run that the cathode should equal 0 volts. Therefore the grid will have to be sourced from the negative supply. A single pot from - suppky to gnd with the wiper bypassed with a cap to the grid should work fine for experimentation. I then placed a low noise current source from - to the cathod and used either a choke on the plate or a resistor to B+.
Remember common grid retains phase so this stage will be out of phase coming off the plate.
Thanks
Gordon
J. Gordon Rankin
Follow Ups:
> Problem is these dacs cannot drive the 100 ohm resistor. Most of the output impedances are like 500 to 1000 ohms. So driving the 100 would be an issue <If they were voltage output devices, yes. But they are current output devices and are designed to put out an exact current WRT the digital signal. They work best with an extremely low load impedance - like zero if at all possible - and many designers using these DACs would say 100R is far too high a load to really provide the ultra low distortion figures such 24bit DACs can offer.
Allen
.
Allen,I was talking to BB last week about the PCM1704 and passive I/V designs. The group of engineers said that with the 1704 or any of their current output devices using very small resistors was a mistake. The problem being the instaneous drive is not capable of delivering to such a load.
The reason for the 2.5K reference resistor in the app was that they wanted to get the best testable results. The output impedance is 1K and too use like a 100 ohm resistor as the load would be too low.
You cannot look at any device in one realm and not the others. Any current source is going to have a voltage of some type and therefore an output resistence. This is I think one of the problems in DIY is that people don't consider all the variables.
Ok back to say the PCM1704. Since the drive range is only 1.2ma it is hardly capable of driving a 100 ohm resistor. The group at BB said that 100 was way too low. Of course they would not give a figure because this is not an acceptable output configuration in their book.
Ok let's look at the AD1865 here too the output impedance is high at 1.7k and the current output at +/-1ma or 2ma range. Many DIY dacs I/V this passively at 680.
Many of the new TDA1543 type dacs want to use low impedance resistors but have found that it comes with multitude of parallel dacs to drive them.
Passive or Passively active (or maybe better said non-feedback) designs have a place where diminishing returns come into play. If the resistor is very small then so is the output voltage. If it is high then the results may not be that great.
The guys at TI/BB thought the idea of using a transformer for i/v to be really stupid. Well until I reminded them that MC cart use transformers for i/v conversion. Of course then the jokes as too who uses turntables anymore. I found some high res picks of some mega$$ tables and emailed them too them and asked them what they thought.
I did mention that they should look to High End for a more unique perspective to what they are making. I don't think anyone is wrong in what they are doing, everyone has their own opinion.
I do agree that the least amount of bs in the feedback path will result in the best output.
My question is why even use feedback?
Thanks
Gordon
J. Gordon Rankin
Gordon,
Now that is something you and I can really agree on 100%!And I offer another 100% - no usage of opamps in an audio signal path - nowhere!
Thoughts on that blunt statement?
And regarding using a traffo for "I/V" conversion - I have two problems:
1/ Most times it is done with a low (100R) load on the DAC and the traffo used for a voltage gain, which i consider missing the point. If it's to be used - then feed the DAC into the low Z side of the traffo directly, and put the load R on the secondary,using it as a current operated device - which is what it is - yes?
2/ It's claimed that the traffo acts as a low pass filter to bandwidth limit any out of band noise from the digital system. Yes it does - but how?
**WARNING - VERY UN SCIENTIFIC TYPE OF STATEMENT FOLLOWING**
I have no science to back this up - but surely if it's HF FILTERING because the magnetic particles simply cannot swop ststes fast enough to pass the HF noise, having them in such a STRESSFUL CONDITION (CONFUSED?) - does this not also limit their ability to carry required microdynamic information within the audio band?
I know that if things are happening just too fast around me I can get very overwhelmed/confused and make bad decisions, should we expect better from tiny magnetic domains with (hopefully) much less inherent computing/reasoning power than a intelligent human?
Allen
Gordon,Allen is absolutely right (and Wright). I think there must have some misunderstanding in your discussion with BB/TI people. The PCM1704 is truly intended to drive a current into a dead short.
Open the link to the data sheet, and go to Figure 5 on page 9. It shows an opamp (an OPA627) with the inverting input connected directly to the Iout pin of the converter. There is no series resistor. Negative feedback forces the inverting input to act like a virtual ground, zero ohms, that the DAC happily dumps current into. The feedback resistor (2.5K you mentioned) converts that current into a voltage on the output of the opamp. But the opamp is really doing the work here. For the DAC to work to full specs, there should be almost no signal voltage on the Iout pins. Don’t let the 1000 ohm output Z throw you; that’s actually pretty low for a current source. In any case the load resistance should be much, much less than 1000 ohms. The BB/TI guys were probably just giving you their standard answer about how the passive approach is not recommended. If you were to load the opamp with a 100 ohm feedback resistor, that would lower voltage output and increase opamp distortion (overload). In a passive use, a 100 ohm shunt load gives a very low voltage (about 25 times lower than what they’d want), and you’d add noise amplifying this back up later; hence the BB/TI remarks. Furthermore, there would then be +/- 0.12 volts of signal swing at Iout that may or may not be too much for the DAC to handle and still meet spec. Now, you may like how it sounds, but it’s not how it was designed to work. There’s no mystery here; this is how current output DACs have been used for decades, and for a lot of applications other than audio too.
PS: Calling an MC step-up transformer an I/V device, while not strictly incorrect, is very misleading.
Open this link for PCM1704 data sheet. See page 9 , figure 5.
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2026, Audio Asylum, all rights reserved ]
