Bal-Bal as I/V
Before I get certain peoples heckles up, I'm just after confirmation that my maths and understanding is correct.
Build idea... miniSharc I2S out 7&8 > TPA COD > Bal-Bal > Sen HD650
...
The TPA COD uses the Ti PCM1794 DAC, a current output DAC needing an I/V stage... current output on each output pin [IoutL+, IoutL-, IoutR+, IoutR-] is 7.8mA p-p
The voltage output level of the I/V stage is 7.8mA p-p x Rf [feedback resistance of I/V converter] ... err... I x R = V
Rf in this case is the feedback resistors of the Bal-Bal [R21, R24, R25, R28]. Furthermore, the input resistors [R22, R23, R26, R27] need to be 0R jumpers...
...
For my maximum SPL, the headphones require 2.2Vrms.
2.2Vrms is 6.226V p-p or 6226mV p-p
Equation time... I x R = V
7.8mA p-p x R = 6226mV p-p
Swap it around ... V / I = R
6226mV p-p / 7.8mA p-p = 798.2... Ohm
So, Rf is 798.2 Ohm to give me my required 2.2Vrms... err!!!
This is where I always fall fowl in my understanding of differential outputs. With 7.8mA p-p x 798.2 Ohm [to 1 d.p] = 6226mV p-p, this is per side [leg, output, phase... whatever it's called]. So, I actually have a output of 12.5V p-p [to 1 d.p] or 4.4Vrms... double what I need... Yes? Correct so far?
So, [again] I really need 1.1Vrms per L/R,+/- output to get my 2.2Vrms.
1.1Vrms is 3.113V p-p or 3113mV p-p
3113mV p-p / 7.8mA p-p = 399.1 Ohm [to 1 d.p]
So, Rf [R21, R24, R25, R28] should be 400 Ohm to give me the 2.2Vrms needed.
*EDIT - I've used the conversion factor of 2.83 for Vrms to Vp-p*
Few... to those in the know, could you give me the nod that the above is right / wrong please.
Kind regards
Paul
Before I get certain peoples heckles up, I'm just after confirmation that my maths and understanding is correct.
Build idea... miniSharc I2S out 7&8 > TPA COD > Bal-Bal > Sen HD650
...
The TPA COD uses the Ti PCM1794 DAC, a current output DAC needing an I/V stage... current output on each output pin [IoutL+, IoutL-, IoutR+, IoutR-] is 7.8mA p-p
The voltage output level of the I/V stage is 7.8mA p-p x Rf [feedback resistance of I/V converter] ... err... I x R = V
Rf in this case is the feedback resistors of the Bal-Bal [R21, R24, R25, R28]. Furthermore, the input resistors [R22, R23, R26, R27] need to be 0R jumpers...
...
For my maximum SPL, the headphones require 2.2Vrms.
2.2Vrms is 6.226V p-p or 6226mV p-p
Equation time... I x R = V
7.8mA p-p x R = 6226mV p-p
Swap it around ... V / I = R
6226mV p-p / 7.8mA p-p = 798.2... Ohm
So, Rf is 798.2 Ohm to give me my required 2.2Vrms... err!!!
This is where I always fall fowl in my understanding of differential outputs. With 7.8mA p-p x 798.2 Ohm [to 1 d.p] = 6226mV p-p, this is per side [leg, output, phase... whatever it's called]. So, I actually have a output of 12.5V p-p [to 1 d.p] or 4.4Vrms... double what I need... Yes? Correct so far?
So, [again] I really need 1.1Vrms per L/R,+/- output to get my 2.2Vrms.
1.1Vrms is 3.113V p-p or 3113mV p-p
3113mV p-p / 7.8mA p-p = 399.1 Ohm [to 1 d.p]
So, Rf [R21, R24, R25, R28] should be 400 Ohm to give me the 2.2Vrms needed.
*EDIT - I've used the conversion factor of 2.83 for Vrms to Vp-p*
Few... to those in the know, could you give me the nod that the above is right / wrong please.
Kind regards
Paul
Last edited:
I've just finished my SE-SE, BAL-BAL and PSU boards.
Everything is fine but there is severe DC-Offset when I use the BAL-BAL in SE-BAL configuration.
Refering to schematics I've connected H55+H53 and H41+H43 to ground the " - terminal"
DC Offset is 2mV (volume turned completely down) to 120mV (volume turned completely up).
What is wrong?
Regards Martin
Everything is fine but there is severe DC-Offset when I use the BAL-BAL in SE-BAL configuration.
Refering to schematics I've connected H55+H53 and H41+H43 to ground the " - terminal"
DC Offset is 2mV (volume turned completely down) to 120mV (volume turned completely up).
What is wrong?
Regards Martin
Quick question!
Anyone knows if the Wire (SE-SE) gets too hot with the new PSU at 15V?
And does the regulator get to hot or can I use slightly shorter heatsinks?
I'm asing becase I'm looking at putting them inside a low profile case, and I want to go as low as possible: toroid will be 24mm high, so the limiting element will be the PSU caps at 33mm (I don't really want to solder them slanted or anything like that, for mechanical stress reasons), with the lowest possible hex spacers.
Thanks in advance for the possible answers!
Anyone knows if the Wire (SE-SE) gets too hot with the new PSU at 15V?
And does the regulator get to hot or can I use slightly shorter heatsinks?
I'm asing becase I'm looking at putting them inside a low profile case, and I want to go as low as possible: toroid will be 24mm high, so the limiting element will be the PSU caps at 33mm (I don't really want to solder them slanted or anything like that, for mechanical stress reasons), with the lowest possible hex spacers.
Thanks in advance for the possible answers!
These numbers do not seem right.
In fact they cannot be right.
OPC would never have designed this much resistance into the feedback circuits, because it would create far too much noise.
Agreed... when you're using the BAL-BAL in true differential mode, it will very effectively cancel out any common mode DC voltage at the input terminals. Say you have +200mV of DC offset at each terminal with a differential input, then the differential output DC voltage will be at or near zero. As soon as you unbalance the input, things change. If you ground the negative terminal in the above example, the input DC voltage is no longer common mode, and since the amp is DC coupled, the DC appearing at the positive terminal will be amplified and seen at the output as a differential DC offset.
You need to make sure both inputs see the same DC resistance. This will require that you solder some input resistors to GND at each amplifier input. If your source has a DC offset, you will probably need a pair of coupling caps and some resistance to GND at each input. Ideally, you should match the DC and AC characteristics of each input.
Regards,
Owen
I'm having a problem with my PSU (v2): I'm using 133kOhm for R30/R55 to give approximately +-12V, however now my output sits at 2V/-20V. Is it possible that this could be caused by a fried C28/54/85/93? Soldering the first one of those took me a very long time and I may have killed it in doing so... I'm getting asymmetrical resistance readings across R30/55 as well as 29 and 52 and the unregulated voltage across the filter electrolytic caps on the input side is the same on both sides, so I'm fairly certain the problem is on the output side.
Last edited:
The edit button seems to have disappeared, so I'm unfortunately forced to double-post here:
I first removed C28/54/85/93, the readings were still asymmetrical. I double-checked R30 and R55, those are both fine. I then removed C55 and C17 and cleaned some flux off and now everything's weird as hell: I'm getting 0 resistance across C28 and 54 and infinite resistance across C93 and 85 - it is my understanding that on both sides, I should be getting something equal to or less than R30+R29/R52+R55 respectively. Is there even any hope left for this board, or should I just start over? I hate abandoning stuff like that, but sometimes it's just necessary...
So, I'm correct in my accessment that trimming 5mm to the heatsinks won't be problematic?
I'm trying to fit my Wire into a Box Enlosures, the less height the better (I have a Gamma2, and I have an urge to keep the enclosures similar)
The heatsink height was a problem, the other one is that I'll have to use really small standoffs (6.5mm probably) and the fact that the aluminium box will function as a big heatsink and will heat up too much.
What are your thoughts?
(Not to mention working with a box that only opens on the front and back)
Last edited:
The regulator heatsinks on my 12v SE-SE build barely get warm so making them 5mm probably wont cause any problems. You could probably get away with the 1in tall version of the BOM heatsink but that wouldn't gain you much room since you still have the tall 3300uf/25v capacitors. Another option is to mount the PSU board on its side
I just played/measured around with it some more and found some more discrepancies:
The tab on one regulator and its corresponding GND pin (#4) are not connected to ground - however, that same side has the short on the actual rail (i.e. across C54/28). I laid a wire from pin 4 to a ground point but apart from the tab and GND pin now being grounded (obviously), nothing changed. Is it possible that I somehow screwed up my PCB with my terrible soldering skills? If so, I'm gonna need to order another one one of these days
the TPS73a01 is a negative regulator. Both schematics for the positive and negative voltages are the same. On the positive side the regulator is in series with the ground. so the ground pin(4) is not connected to ground on the positive output but connected to the positive rail.
You must also be using a transformer with 2 secondary windings and not a center tapped winding. When comparing voltages on the 2 regulators you cannot use the ground terminal on the circuit board as a referance point for the positive regulator you must use pin 4 ie the positive output. Are you getting the require reference voltage 1.175V across the 105k resistors r55 r30?
You must also be using a transformer with 2 secondary windings and not a center tapped winding. When comparing voltages on the 2 regulators you cannot use the ground terminal on the circuit board as a referance point for the positive regulator you must use pin 4 ie the positive output. Are you getting the require reference voltage 1.175V across the 105k resistors r55 r30?
Damn, I spent so much time looking at the schematic and didn't notice those things... But still, the fact that there's a short between +12 and GND just can't be right. I've started cutting into the PCB to cut any possible shorts but that isn't helping either. I think I'll just order a new PCB and start over. A blown regulator (although I wouldn't know what blew it) might also be at fault..
Hi Paul,
Looks like your question has been missed
I haven't had time to check it in detail, but worth checking with the ivy iii manual for cod and opa1632 as I/V.