GroupDIY
General Discussions => Drawing Board => Topic started by: Spencerleehorton on November 10, 2020, 02:11:37 AM
-
Hi,
Just thought I'd document my api-312 500 series build.
Single sided pcb which you can make through the iron on method using laser printer, winding my own txfs, 48v, -10db pad, polarity switch, di input, gnd lift, output attenuator pot.
Gonna use a couple of 2520 op amps I had left over.
Have design and printed pcbs and going to put them in ferric chloride then drill.
Will detail parts etc once I get to that stage.
-
A 10dB pad is hard to do whilst maintaining the right loads for both the microphone and the transformer. 15dB is about the minimum you can do and meet these criteria, 20dB is more normal.
Cheers
Ian
-
I wondered that Ian, I might just do the pad the normal way rather than this capacitor way.
Normal way I do it is with 3 resistors, 120R and 2 x 470R.
-
I wondered that Ian, I might just do the pad the normal way rather than this capacitor way.
Normal way I do it is with 3 resistors, 120R and 2 x 470R.
That should be OK. I use 150R and 2 x 680R
Cheers
Ian
-
oh right, so i could have, 120R, 470R or 150R, 680R or 160R, 750R.
what would be the differnce between them?
i actually find that with my api-312 preamps at the moment my -20db pad is a little bit too much, as in, when i pad for instance the snare without it padded its a little bit too much and when its padded i have to crank the gain nearly all the way!!! think im using 120R and 470R.
-
oh right, so i could have, 120R, 470R or 150R, 680R or 160R, 750R.
what would be the differnce between them?
Typical mic pre designs are arranged so their input impedance is about 10 times the expected source impedance. Most people assume the source impedance is typically 150 ohms so the input is designed to look like 1500 ohms. To maintain this with a pad in circuit you need to do two things. First arrange the output impedance of the pad to be about 150 ohms - which is the value of the padding resistor I use, and also for the input impedance to still look like 1500 ohms which means the sum of the two input resistors needs to be at least 1500 - 150 =1350 ohms. A pair of 680 ohm resistors will be 1360 ohms.
Of course it is a little more complex than that, but there is quite a bit of leeway, so lots of combinations in this region will do the job.
i actually find that with my api-312 preamps at the moment my -20db pad is a little bit too much, as in, when i pad for instance the snare without it padded its a little bit too much and when its padded i have to crank the gain nearly all the way!!! think im using 120R and 470R.
Depends on the gain range of your preamps but it should not be an issue in practice. If your gain range extends from 20dB to 60dB it should not be a problem unless your pad is a lot more than 20dB.
Cheers
Ian
-
Ah ok I'll stick with your values, 150R and 680R and see how that works.
-
good luck for your build, please keep us informed and post some pictures of the process
best regards
-
No probs, I'll post some pics of the pcbs later once drilled.
Then I need to get the components together and get winding!!
-
Thought i'd share the schematic im working of to see what you guys thought, im not going to have the -10db pad on the instrument input, going to have -20db on the mic input and polarity and 48v added.
Will leave out the gnd lift probably and add the 300R resistors for the 1k log pot output off the secondarys of the output txf.
Just interested in whats going of with this arrangement if anyones got any thoughts?
-
Spencer, please allow some remarks,
your "22kOhm log pot" will be a rev log taper pot,
the +/- orientation of C3 is depending on type of DOA used,
125uF is a last century parts value for C1 and C2 (btw with original parts value 12.5uF) and fitting a 35VDC capacitor rating will allow the use of +/-24VDC supply rails as well at assumed same cost,
zobel parts values of R3 and C5, if needed at all, is depending on type and series or parallel hookup of T1,
J2 pin1 connects to shield, not ground (different symbol/node), that is not neccessarily the same in this spot,
C6 and C7 will be placed close to XLR-in, maybe not on pcb, caps common side connected to shield, not ground (in order to more effectively block/reduce rf coming in from the outside world, fi no mobile phones back then...)
best regards
-
Harpo,
Just to qiery your term "shield" do you mean 0v/audio GND rather than chassis GND?
also i will use 470uf/35v for C1/C2.
I'll update the schematic top what i think you've described.
-
here is the updated version
-
Ive changed it again, bit happier with this one.
what do you think?
-
Just to qiery your term "shield" do you mean 0v/audio GND rather than chassis GND?
Shield meaning chassis GND. This node, as well as phantom current return, is connected to 0V/audio GND in a single spot, probably at the psu.
also i will use 470uf/35v for C1/C2.
A heavy rise of capacitance (factor 37.6). Depending on channel count, is your psu able to charge these?
-
Yah my psu is pretty hefty and it's only going to be two channels.
-
So I'd be better with 125uf/35v for c1/C2?
-
Iirc API used 1k pot at the output without 300R resistors, it works well for other trafo coupled line outs too. Haven't compared both and wonder if there is any change in sound between them.
-
It's a good talking point defo the use 300R with the 1k, if anyone could explain?
-
fwiw, smaller psu caps are typical. I imagine the Forsell 990 has the additional two 3u's for sake of transients.
(https://i.imgur.com/12DzXen.jpg?1)
Also your 220 output cap could likely be raised to 16v without issue.
the use 300R with the 1k, if anyone could explain?
Someone correct me but I think the reason for the 300r's is for sake of connecting to older gear. Also at full CCW, the 2ndary would otherwise reflect an unloaded short (not sure how much it matters).
-
Ah ok I'll go for the 47uf/35v then, it's just in my access 312 pcbs it has 470uf in those positions.
Perhaps I should take them out and replace!!
-
i've put in a 47uf/35v and a 68uf/63v as this is what i have on hand, will order some more and replace along with some 12v zeners, only have 11v and 16v zeners
-
cant seem to find a schematic for the 25k rev log stepped gainswitch?
could anyone post please? as i want to make a couple form these lorlins
this is what ive guessed at:
1. 15k
2. 4.7k
3. 2k
4. 1.2k
5. 750R
6. 510R
7. 360R
8. 240R
9. 180R
10. 120R
11. 82R
does that seem about right?
Edit: these values are pretty close and should work
-
Perhaps I should take them out and replace
Should be fine as is. Compare to be sure.
Some helpful advice and links on stepped attenuators:
https://groupdiy.com/index.php?topic=44800.0 (https://groupdiy.com/index.php?topic=44800.0)
-
Just waiting for 120pf and 220pf styros, have put in some temporary txf that should get signal through until I come up with a winding plan for the output txf.
Building the psu today and hopefully testing with some power as I can just pop in some poly 120pf and 220pf to get it working.
Will post some pics soon.
-
well it seems there are many errors in that schematic so im following the api version now as it all seems correct.
1st channel i've powered up and had a slight fault on the instrument input which was just a dodgy connection to gnd.
i'm running at 16v at the moment, whats the advantages of running at 18v? just curious.
working on the 2nd channel now and the case, need to drill a few holes etc and mount the pcbs in there.
also going to wire the polarity but i notice the api VP312 version has the polarity before the output txf as with the access 312 versions they have the polarity before the input txf? does this really matter? is there a preference?
will wire in the 48v and pad switch as per the VP312.
i have noticed on the VP312 shematic R3 is 10k, where i have used 5.1k here so will change and see if anything changes?
stepped switched work great with lorlins.
-
A few pics
-
Insides
-
And yes I get that it's not 500 series yet, but I havent managed to get into my shed to do metal work.
Just managed to get hold of some new brass to make some u47/fet47 headbaskets, so got a couple of weeks next month were I'll hopefully knock out a few mics and do the metal work.
At the moment to get them working using this 2U enclosure which I was going to put 2 x neve 1290 in.
-
i'm running at 16v at the moment, whats the advantages of running at 18v? just curious.
marginal higher headroom, if needed at all, or some smoke if your type of DOA isn't designed to take this amount of +/-18VDC (=36VDC differential between rails) supply voltage.
also going to wire the polarity but i notice the api VP312 version has the polarity before the output txf as with the access 312 versions they have the polarity before the input txf? does this really matter? is there a preference?
with polarity flip before the output transformer, you could change polarity of the line input as well.
i have noticed on the VP312 shematic R3 is 10k, where i have used 5.1k here so will change and see if anything changes?
already told you, this resistor is part of the zobel network and its parts value is depending on type of transformer and if you hook up the primary windings in series or in parallel.
stepped switched work great with lorlins.
nice. My take on this subject would be based on some calculation for a linear dB increase, like this excel sheet (http://www.hausverwaltung-heger.de/al_leck_trick/API-312_SwitchStepCalc_Spencerleehorton.xls). Just overwrite the parts values or number of step positions in the yellow cells to match your plan.
best regards, Harpo
-
after a few mishaps with volkers PSU, i blew a few LM317, a 240R got very hot and a 100uf cap exploded!!
found my errors, 50k pot rather than 10k, changed the other 10k pots to make sure and replaced all the other bits, works like a charm!!
Both channels now working, just need to wind another input txf for channel 1 as its not looking pretty!!!
Now onto the 48v, Pad and the polarity, nearly done but i do keep smelling something getting hot?
need to keep an eye on these 2520 opamps!!!????
-
need to keep an eye on these 2520 opamps!!!?
Not if you added something like 10ohm 'protection' resistors to P-in.
BJT's 139 and 140 should get warm but not blazing hot.
-
I've gotta laser heat thermometer and will go over the circuit and see where the smell is coming from!!
-
channel 2, BD139 and BD140 are getting very hot!! ??? im using Gar2520s.
will have a poke about and see what gives!
-
channel 2, BD139 and BD140 are getting very hot!! ??? im using Gar2520s.
will have a poke about and see what gives!
Either an assembly error or the bias is set too high.
Cheers
Ian
-
also going to wire the polarity but i notice the api VP312 version has the polarity before the output txf as with the access 312 versions they have the polarity before the input txf? does this really matter? is there a preference?
Switching at the input can cause loud pops. I always put polarity reverse at the output.
-
I haven't had a chance to poke around yet on ch2 but I'm only feeding 16v in at the moment and it takes a few minutes for the bd139/40 to get hot.
Ian, you say bias could be set too high?
Do you mean on the 2520 or elsewhere?
I dont have a bias pot or anything?
Yes, I'm going with polarity before output.
Will wind input txf this morning, change it over then check ch2 and compare voltages to ch1, hopefully will find a little mistake!
-
I haven't had a chance to poke around yet on ch2 but I'm only feeding 16v in at the moment and it takes a few minutes for the bd139/40 to get hot.
Ian, you say bias could be set too high?
Do you mean on the 2520 or elsewhere?
I assume the BD139/140 are on the 2520? So if they are BOTH getting hot then they are passing too much dc current. Did you build the 2520 yourself? Maybe that is where the error is.
Cheers
IAn
-
if they are BOTH getting hot then they are passing too much dc current.
Both channels now working
What does idle output or 1kHz look like on scope? You mentioned blowing a few 317's. Might have affected 337 side.
-
On first look I have a 68uf/35v and a 47uf/35v in c1 and c2, this might be my problem!!
Will change both for 47uf/35v
16v is good on psu on +/- and 48v is good, so no problems there
-
Sorry, my ch1 has 100uf/63v for the first caps after the diode for power coming into the 2520.
I will change ch2 to the same and see if still get the same problem.
-
Hi Spencer. I just clicked on the link to your website and got an error message. It says "There has been a critical error on your website." followed by a link to 'Learn more about debugging in Wordpress'
Cheers
Ian
-
Yeah my website is a bit up the spout, if you reload it, it does come up?
Need to rebuild it really
Both channels bd139, bd140 are both getting scolding hot after about 3-4 minutes with 100uf in?
Just wound another input txf, tested it, checks out fine, no shorts, dcr good, inductance great.
Put in channel 1 and I get cut off waveform and when I adjust output level, all the way up I see waveform cut off top, all the way down I see top part of waveform but no bottom!!!?
-
Yeah my website is a bit up the spout, if you reload it, it does come up?
Yes, that does it. No idea why but it does seem to work.
Cheers
Ian
-
just realised i still have in 5.1k and meant to change to 10k for the R3 resistor, could that be why its drawing too much?
-
Found a couple of errors and changed, now the bd139/40 just seem to get warm after about 10 minutes, can still hold my fingers on them and not get burnt like before!!
D5 wasnt down to gnd and the 10k/220pf were the wrong way round.
Ch2 looks good and I think I can continue and put 48v in and polarity and pad switches and finish this channel.
Ch1 waveform is not looking good, I've used the same spec output txf as ch2 but wound my own input txf.
Normally my input txf come out fine, but will swap with ch2 input txf and see if waveform changes, if it doesnt the problem lies somewhere else?
-
I will check where the waveform gets weird as I suspect its after the 2520.
-
Changed over 2520 for another I had and waveform is good now on both channels.
Will finish off other switches then give these channels some use!!
-
Well it seems ive hit a weird one again!!
both channels work fine through instrument DI input, get nice level and with gain can boost and pad works.
On mic input i get very low level at output with gain on full?
ive wound 1:1:3.5 ratio txf and have Primaries parallelled so i get 1:7 step up and just dont get the output?
Ive tried different output txf but the level doesnt seem to be coming out of the 2520?
am i missing something here?
I have C1 = 220pf and R3 = 10k with R2 = 150k, R8/R10 = 680R and R9 = 150R
R4=20k with C2 = 20pf, R11=200R and C8 = 220uf.
I have left out R12(1R)
I have both C6 and C5 = 1000pf
48v is working fine (R5=200R with R6/R7 = 6.8k, R1=12k with also R13= 10R/5W to chassis.
something must be loading it down but so far i havent spotted it!!????
-
a couple of errors to fix, the jack socket was putting the 0v to chassis so ive put in some isolated jack sockets and C2 (100uf/63v neg leg was down to V+!! so have moved it to GND!)
have taken off the zobel but will put back on 5.1k and 220pf and see what the rings like!
-
I've wound some output txf for these, they are on slightly bigger bobbins and lams.
250 turns of 0.25mm x4 and I end up with 11 ohms and 130mH inductance per winding, I even managed to twist the four winds!!!
This tests very close to my fab 2503 versions which are 10R and 104mH.
Input txf wise I've wound a 1:8 to the same spec as the 2622.
Most of my problems have been due to the 2520 I've used and out of 4 I had left over they all seem to have problems do I've ordered two more which should hopefully fix issues.
Zobel ive ended up with 10k / 220pf and 10k to gnd.
Gain switch I have 20k / 100pf and 22k stepped with 220uf cap to gnd.
C1/C2 is 47uf/35v with 1n4007 diodes.
Only issues now is the 1k log attenuator seems to adjust output on pin1 but not on pin 3? And when flip phase I get different levels on pin 2 and 3?
I think once I change 2520 it will all be fine.