dc output control loop

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
im sure its a stupid question but i gotta ask cause i cant find my answer anywhere.
for the ina217 they show the optional dcout control loop on the schematic. using an opa237.
so my qwuestion is what pins would be used on the opa237?
and can i use any opamp?
and once i know this i will never some of your brain to to look over my preamp design
ina217micpre.jpg
 
So ... a balanced input to unbalanced output circuit. I love it. I've been looking for a source for a board like this for months. This will make a nice front end for my aussieamplifiers.com NX500 amps.

FYI: The OPA137 see: http://focus.ti.com/lit/ds/symlink/opa137.pdf ... pin connects are industry standard for single channel op-amp = mini DIP 8 pin = 2, 3, -/+ input, 4 & 7 power, 1, 5 & 8 = NC ... but you already knew that or you would not have been able to complete the board design.

Evaluation:

> schematic = no problems found.
* The use of the opa137 is optional, but will improve the quality of resulting audio in many cases.

* The requirement for phantom power only applies to microphones. This may not be necessary, in which case a lot of the front end parts, etc. , can be left off of the boards during assembly, but it is a good idea to keep the traces & through holes for this ... jus' in case. (My usage would be for interconnect between multi-channel pre-amps and bi-amped speakers, XLR cabling being shielded and better CMRR on longer lines ... the usual reasons for using pro-audio stuff instead of RCA type cabling.)

> Circuit board layout suggestions.
* Open up the gap between the traces somewhat. During DIY assembly, solder can axidently spill across narrow gaps ... not that you are that sloppy, but I am. Gap might be as large as 1/16" or even more. Common practice a couple of decades ago was a gap of 1/8" so that us ham fisted gringos could assembly 'em without too many errors, especially around the solder pads and through holes. (A lot of us old guys prefer British Units (inches = ") rather than the French hoseup ... 'cause we are peculiar and appreciate a True Earth Centric measurement standard than a fictitious one.)

* The mount holes on the four corners should be of a compatible diameter with the computer motherboard plastic standoffs = those cute little plastic push ons with the gummy backs are cheap and plentiful and available at all most all surplus stores. And make sure there is plenty of gap around the mount holes for the same reasons mentioned = sloppy old geeks acidently using metal standoffs and grounding out the back planes & traces, etc.

* It looks like there is room around the DC blocking caps for some extra through holes. This will allow for the use of bigger bi-polar electrolytics or even MKT type blocking caps. (BP1 and BP2)
===0=0=0 0=== instead of =====0 0==== ... or vise versa ... same, same around other parts that may be optional (using jumpers) or of alternate physical sizes or types. (Around E$17 and E$18 = are those coils? ... I wouldn't want 'em and would substitute for jumpers, others might want a different size, etc. ... it is advisable = greater versitality / usefulness of the basic board layout.

* The Ground Plane (connected to pin 1 of XLR) is a little overpowering = too much copper left behind. Keep the gap between the Ground Plane areas and the actual signal paths bigger as well ( Pin 2 & pin 3 of XLR input thru the DC blocking caps and on thru to the imputs of the first stage (INA217 op-amp) ... and then on thru the output and around the balancing op-amp input & output traces (OPA173)

* ?? Is that an extra line driver I see? (DVR134) ... If so, putting this extra set of parts, solder pads & through holes on this board might be a waste, somewhat, for the purposes here = balanced to unbalanced and back to balanced is not the best idea for quality audio "pass through" or daisy chain of XLR connections. Too many op-amps spoil the result as it were = each amp stage compounds the distortion and errors of the previous stage = possibly twice the over all distortion and error even without any gain. ... I see where you want to go with this = an XLR pass through "breakout box" with an unbalanced tap ... Much has been written about these kinds of efforts. All electronic taps have a problem of one kind or another: TV splitters = loss of signal strength, digital splitters = timing consequences & reduced performance, analog audio breakout boxes like this contribute to distortion or depreciation of the noise floor or degradation of some aspect of quality. Leave the holes, prepare to ignore the usage.

* Follow the signal ground connections around the circuits' op-amps and passive components, looking for any topography other than a "star" like grounding (+15 / 0 / -15 ... 0 = signal ground). It appears that you understand the differences between signal ground and the Ground Plane that is attached to the XLR shield (pin #1) ... inattention to this and the topograhpical layout of the signal paths (+, -, gnd, shld = balanced and plus, minus as shield = balanced) ...

Keep the faith. This is a worthwhile project and I will buy some of your boards or help defray some of you costs when you are ready. (I would offer to buy a couple of these, as is, but I would leave a lot of the parts off and maybe even chop it up with a saw ... :D )
 
fast eddy,
thanks for your input. i'll get back to the drawing board.
What do you think about the option of balanced and unbalanced outputs? I agree as far as two many op-amps will degrade the signal but i like the idea of having balanced outs if needed. and i like what i've heard and read about the drv134 and its ability to boost level 6dB.
 
Fast Eddy
ok i took your advice and i believe i covered most or all aspects but like i said this is my 1st layout so hopefully you or someone would look it over.
im going to work on the power supply and then i will join the boards. I assume this would make it more cost effective. and basically like the Audiosector boards, how they snap apart.

Also because i will use this for a mic i would like to have a VU meter or a clip light. Im not to sure how i should rig this up or where. anyone know of any links.

Also i was thinking after seeing this idea somewhere, to put a patch point somewhere in between the DRV134 and the INA217 so if i decide later to make say a compressor i could put it all in a single RU.
anyways here it is
pre1-1.jpg
 
" ... What do you think about the option of balanced and unbalanced outputs? I agree as far as two many op-amps will degrade the signal but i like the idea of having balanced outs if needed. ..."

I like it too, but I would take a look at how it is done on a pro amplifier ... usually done with passive elements rather than active op-amps. Checkout a Crown Amp with the XLR pass through ports = usually done with transformers and/or an array of other passive resistros, etc.

It is desirable to have a pass through XLR port for general purposes, but for higher quality audio work, the simpler balanced in to unbalanced out is better. It is possible to add the XLR pass through ports on another board, so having some through holes for this accomodation may be in order if you decide to remove the rest = very optional as I can live with the parts' spaces / holes on the current board design.

As for your latest board design: Consider removing that "tongue" of copper between the trim pot traces = this encourages crosstalk between the signal paths and the Ground Plane ... same, same with the Ground Plane "tongue" that parallels the signal path 'tween the blocking caps' traces, BP1 & 2, and the first stage op-amp inputs ... same, same with the Ground Plane island 'tween and around the clamping diodes, etc. ... and there are several other Ground Plane "tongues" and islands that might or might not degrade the audio signal. It is common practice to have these "tongues" and islands on power supply boards, but modern "audiophile", golden ear practice and theory of design of amplifier boards indicate that this is just another source of "stray capacitance" and possible crosstalk at higher frequencies ... those parallel ground "tongues" and signal path traces acting like tiny capacitors, bleeding off the signal quality ...

I see you opened up the gaps 'tween traces somewhat ... IMOP these could use some more 'tween and around any signal path = reducing the above mentioned stray capacitance, but if you remove and reroute those Ground Plane traces as above, this may not be required or even a possible bother. ... Fat Ground Plane(s) around the perimeter of an audio board = very good = act as shielding. Ground Plane traces in and amoungst the signal traces = not so good and in a few cases, bad .... :smash:
 
" ... so you think i need transformer outs? ..."

I did not mean to interfer with your design perameters ... so I would suggest you not significantly change your existing approach and try for transformers as "pass through" output to XLR ... but it would not hurt to have solder pads & thru holes ready as an option for other users of you boards & design. ... In fact leaving the output impenedence matching op-amp connections AND transformer output connections on the board ... both as options ...

Personally, I will just not install those parts on any boards I might use, bypassing that need, since I would be simply interconnecting between balance XLR input connectors to an unbalanced input amplifier module. ( adding them to this type of amp: http://3dotaudio.com/ampics.html )

As for the costs of XLR I/O transformers ... you get what you pay for. There may actually be a very good qaulity balanced to unbalanced PCB mount transformer (I've seen 'em) but I don't have a link or suggestion for this.
 
fasteddy, here is another revision. tips?
i noticed on the drv134 i had V+,V- pins crossed up. it gave me more of a worry then it should've thinking that i woouldnt be able to keep the board single sided. anyWAYS where should i have the taps for transformer out? is it fine the way i have it set up? im not too familiar with that process i need to do some research.
An externally hosted image should be here but it was not working when we last tested it.
 
impsick:

1) Still wondering what those inductors accomplish on the input of the XLR ... ??? ... a band pass blocker / filter? ... don't you trust the '217 op-amp to filter off the EMF ? By reducing the blocking caps to around 4.7 uF accomplishes approximately the same thing = HF noise reduction ... (Personally I will probably just put jumpers there and go with your suggested 47 uF blockers and trust metal case shielding to take care of the High Freq noise ... or experiment, now that there are some extra solder pads there to work with = :>) )

2) The clamping diodes, + rail to signal path(s) to - rail = you might consider using some Schottky diodes here instead or ordinary silicon diodes ( http://en.wikipedia.org/wiki/Schottky_diode = very fast recovery, fast response) ... this also reduces any incoming EMF noise on the inputs and of course clamps the incoming signal to the rail voltage plus 'tween 0.2 and 0.4 volts (instead of 0.6 for silicon).

3) Good show on the 220 uF filter caps ... I'll probably use Tantalum here and leave the electrolitycs for the off board power supply. (Because there are several solder pads to chose from, I will be able to do this without fit problems.) The snubbing caps (0.1 uF) should be plastic (polystyrene) in any case, your circuit or mine = obviates any preceived need for Tantalum or other. (This snubbing cap trick is what Analog Devices does to make their op-amps look good on the spec sheets and appliucations guides ... :cool:

4) Incidentals and my build notes:

... Missing mounting holes ??? = size 'em for those cute little gum on plastic standoffs found on computer motherboards ... extra mue bweno

... Find some place to put your copyright notice = looks real good on your resume'. ... ;)

... I will probably leave the '132 chip off, but keep a mini-dip socket in its place at that point for my unbalanced output ( may also cut the trace to chassis ground (close to the '132 pin 4) just below that DIP socket and jumper a resistor ro the chassis ground ... or not and just use it as is ... I'll get my meters and check it out. (Usually I like to have unbalanced signal ground separate from the chassis ground plane ... you never know who or what might touch the case and blow the amp(s) downstream, plus this can be a source for gound loops and other noise. ... complete isolation of the ground plane from signal in and signal out = isolation from from those pesky circular currents in the copper ground plane, etc. ... I use # 18 or #20 twisted pair stranded silver wire, teflon coated with a copper web shield and teflon jacket on both balanced XLR connections and unbalanced connections ... tying the shields to that common ground plane, but leaving the unbalanced (-) black lead (signal ground, not chassis ground) to connect to an independant point very close to the output of the op-amp ('217) ... refering to your original diagram:
ina217micpre.jpg
... I might chose the point between the two trim resistors R8 and R7) and add a cap there to the chassis ground = stabler output adjustment and a well founded ground ... BUT, don't let me change your basic design and layout = I want a dozen of 'em ... that's right, I want to buy 12 boards when you get ready to have 'em made ... email me.

:smash:
 
FastEddy
sweet man thats cool, i'll get in touch with you in the next few days. I'm just gonna fix a few more areas and see your final word before we get them made:devilr:
oh the inductors are there because as stated in the data sheet by puting an input network consisting of a 1.2uh, and 47k resister could greatly reduce oscillation when a very low impedance source is connected. so its just there on for the safe side. of course jumpers will always do if not needed. Thanks again!
 
That last layout image I saw seemed to have the +48V filter cap (NP3 ?) in series with the +48V supply. That's not what the schematic shows, but what the layout appears to show. Needless to say, if it's really laid out that way you won't ever get phantom power.
 
" ... the inductors are there because as stated in the data sheet by puting an input network consisting of a 1.2uh, and 47k resister could greatly reduce oscillation when a very low impedance source is connected. ..."

Like a tube pre-amp ?? = output impedence can be / usually is around 600 or 700 ohms or so, but some can get quite tricky and they love an inductive load, so, of course leave the solder pads and traces in place ... (Example: http://www.bottlehead.com/et/adobespc/foreplay3/foreplay3.htm ... I have one of these coming, but may rely on the device to limit ocsillatory questions.)

rco3: " ... That last layout image I saw seemed to have the +48V filter cap (NP3 ?) in series with the +48V supply. That's not what the schematic shows ... "

I saw that too, but since I'm not interested in the phantom power (just yet), I didn't mention it = good eyes.

:smash:
 
hey eddie whats up man? ok i think this is it. if you could i'd appreciate it if you looked over it one last time. i want this thing to work to its fullest so if you have any suggestions please tell me. thanks
An externally hosted image should be here but it was not working when we last tested it.



"balanced XLR connections and unbalanced connections ... tying the shields to that common ground plane, but leaving the unbalanced (-) black lead (signal ground, not chassis ground) to connect to an independant point very close to the output of the op-amp ('217) ... refering to your original diagram"

can you explain what you mean by this?
 
" ... leaving the unbalanced (-) black lead (signal ground, not chassis ground) to connect to an independant point very close to the output of the op-amp ('217) ..." ... "can you explain what you mean by this? ..."


Long winded: Well, common audiophile golden ear practice is to have a single spot on a circuit board dedicated to signal grounds and power bus (rails) ground = thus reducing a lot of ground loop possibilities ... OR ... having the output signals of any op-amp stage (in this case, the second stage op-amp) with its own ground point, but not connected to the massive ground plane around the edges ...

This reduces the so called "eddy currents" in the copper from interfering with or otherwise mucking up the signal with extra noise, potentially, in the form of stray capacitance and cross talk from other areas on the board. This is also one of the reasons that op-amps are biased to very little or no DC offset on their outputs = any out of bounds DC voltage flows back through the other parts and eventually winds up pushing on the common (shield) ground plane causing noise or part failure. (Hydraulic model = "pushing" = stray voltage potential.)

Short version: isolated signal output grounds tied to the power supply (+/gnd/-) ground at a single central point, not the outside ground plane = quieter operation, reduced ground loops. :smash:
 
FastEddie i got a question for you.

ok so i had made a pcb of the original design a few weeks back and i got to testing it the other day. So its not the best of layouts, and instead of the BP-47uf caps on the input, i used some Muse BP 100uf's cause thats all i had. So this is what happened.... I pluged in a Audio-Technica AT2020 condensor mic and any spl in the general direction of the diaphram and it would clip. (hard)

So after awhile i then pluged in a SuperLux CM-H8a and to my surprise it was fine. Nothing clipped or distorted.

I checked the AT2020 with another phantom power source and it was fine. But back in the DIYpre again everything was too hot for it.

Any idea of what this could be or what i can do to troubleshoot? thanks, Omar
 
" ... So its not the best of layouts, and instead of the BP-47uf caps on the input, i used some Muse BP 100uf's cause thats all i had. ..."

Of course we should see the complete "as built" schematic as opposed to the original plan.

" ... and it would clip. (hard) ... "

My guess is that for some reason the gain on the XLR input is way up there = outta control ... OR ... somehow the phantom power parts of the input are driving through the caps and injecting some DC component or dramatic signal into the first op-amp input(s) ... this 2nd question is the more logical for "hard clip" to the PSU rail voltage ... stick a voltmeter right on the op-amp input and check for DC ... any greater than a few millivolts = bad phantom power circuit.

Another possible is that you took my advise and used schottky diodes for input protection and somehow got some that are defective (Oops!) = use ordinary silicon diodes for now (like 1N4001 type, etc.) ... FYI: a "hard clip" could indicate the phantom power is running away (as above), so using a healthy silicon diode here protects the op-amp(s) ... also look for heat in the phantom power circuit.

You might try a simple signal generator instead of a phantom powered mic (and temporarily remove the phantom power parts from the board) = the input would then be under your control.

.... Next consider something similar happening downstream = output to rail voltage on the downstream op-amp(s) ... and possibly some input connection to a rail voltage source.

Of course you checked for the obvious = shorted traces during solder connections / construction, etc. ... right?

Re: http://i59.photobucket.com/albums/g307/impsick/ina217micpre.jpg ...

Note that I start at the input and progress to the output looking for "out of tune" op-amps ... then backtrack to other possibilities ... my diagnostic methods may be opposite of others = some would start at the output and work back upstream ... since you report no popped op-amps (smoke), I assumed that the very first stage was the question ... :smash:
 
Ya your right about needed to check this with the best layout not the previous. I had jumped ahead of myself and etched a board and decided to try it since i had it. plus it should help improve even my final layout.
I used the IN4148's instead, cause that was what i had around.
i will do as you suggested and try a few things. Although i dont understand why the other mic responded without cliping.?

anyways thanks for your swift reply. You off today or at work, either way take it easy, talk to you later

peace
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.