API 312 Thread!

Shop deals on ebay
Advertise with us
shop deals on amazon
GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
I'm packaging an API 312 but I'm having a bit of a difficult time wrapping my head around the stepped attenuator calculations. I have done some reading and I've tried a few of the online calculators.

What I want is a 20 step attenuator that approximates a 22k rev log curve, with each step being about 2db. Much like the 12 step that folks here are building on the DIY boards. In my case I'm using a vintage 312-5 card and I already have a high quality shorting 20 step switch. I will be using a switchable 20db input pad, so that should give me a reasonable range of gain control.

When I run a calculator, it does not seem to relate to the 12 step resistor values on the DIY circuit boards.

If I use this, entering the values 22,000 (total value) and 20 (steps):

http://homepages.tcp.co.uk/~nroberts/atten.html

Results:
<<<<<<<<<<<<
Step 1, Attenuation = Infinity dB, Rx = 22000 ohms, Ry = 0 ohms, Resistor = 0 ohms.
Step 2, Attenuation = 75 dB, Rx = 21996 ohms, Ry = 4 ohms, Resistor = 4 ohms.
Step 3, Attenuation = 69 dB, Rx = 21992 ohms, Ry = 8 ohms, Resistor = 4 ohms.
Step 4, Attenuation = 63 dB, Rx = 21984 ohms, Ry = 16 ohms, Resistor = 8 ohms.
Step 5, Attenuation = 58 dB, Rx = 21972 ohms, Ry = 28 ohms, Resistor = 12 ohms.
Step 6, Attenuation = 53 dB, Rx = 21951 ohms, Ry = 49 ohms, Resistor = 21 ohms.
Step 7, Attenuation = 48 dB, Rx = 21912 ohms, Ry = 88 ohms, Resistor = 39 ohms.
Step 8, Attenuation = 43 dB, Rx = 21844 ohms, Ry = 156 ohms, Resistor = 68 ohms.
Step 9, Attenuation = 39 dB, Rx = 21753 ohms, Ry = 247 ohms, Resistor = 91 ohms.
Step 10, Attenuation = 35 dB, Rx = 21609 ohms, Ry = 391 ohms, Resistor = 144 ohms.
Step 11, Attenuation = 32 dB, Rx = 21447 ohms, Ry = 553 ohms, Resistor = 162 ohms.
Step 12, Attenuation = 29 dB, Rx = 21219 ohms, Ry = 781 ohms, Resistor = 228 ohms.
Step 13, Attenuation = 26 dB, Rx = 20897 ohms, Ry = 1103 ohms, Resistor = 322 ohms.
Step 14, Attenuation = 23 dB, Rx = 20443 ohms, Ry = 1557 ohms, Resistor = 454 ohms.
Step 15, Attenuation = 20 dB, Rx = 19800 ohms, Ry = 2200 ohms, Resistor = 643 ohms.
Step 16, Attenuation = 17 dB, Rx = 18892 ohms, Ry = 3108 ohms, Resistor = 908 ohms.
Step 17, Attenuation = 14 dB, Rx = 17610 ohms, Ry = 4390 ohms, Resistor = 1282 ohms.
Step 18, Attenuation = 12 dB, Rx = 16474 ohms, Ry = 5526 ohms, Resistor = 1136 ohms.
Step 19, Attenuation = 10 dB, Rx = 15043 ohms, Ry = 6957 ohms, Resistor = 1431 ohms.
Step 20, Attenuation = 8 dB, Rx = 13242 ohms, Ry = 8758 ohms, Resistor = 1801 ohms.
NOTE:
An input series resistor of 13242 ohms is required to provide the initial 8 dB of attenuation.
>>>>>>>>>>>>

I'm not sure how to implement these results. Should I be using his shunt calculator instead?

http://homepages.tcp.co.uk/~nroberts/shunt.html

Also, when I attempt to use the following calculator the results appear in my browser window behind some of the entry fields, making them illegible. I've tried several different browsers but the results are always the same.

http://www.quadesl.com/attenuator.html

My expectation is to come up with a list of resistors similar to the values used in the 12 step attenuator, but with a total of 20 steps. I believe the 12 step values are:
20k
10.8k
6.5k
4.12k
2.68k
1.75k
1.15k
730
450
260
120
0

Finally, just to make sure I don't assemble this thing backwards-- am I correct in assuming that higher resistance at the attenuator switch results in less overall gain?
 
[quote author="ultrasonic"]I'm packaging an API 312 but I'm having a bit of a difficult time wrapping my head around the stepped attenuator calculations. I have done some reading and I've tried a few of the online calculators.

What I want is a 20 step attenuator that approximates a 22k rev log curve, with each step being about 2db. Much like the 12 step that folks here are building on the DIY boards. In my case I'm using a vintage 312-5 card and I already have a high quality shorting 20 step switch. I will be using a switchable 20db input pad, so that should give me a reasonable range of gain control.

When I run a calculator, it does not seem to relate to the 12 step resistor values on the DIY circuit boards.

If I use this, entering the values 22,000 (total value) and 20 (steps):
...........snip
[/quote]

Hi ultrasonic,

this is not an attenuator switch. It is part of a noninverting op-amp,
with voltage gain av=1+(Rf / Rx), where Rf is your fixed feedback resistor R1=22kohm.
Rx varies from R2=200ohms to R2+your rev log value of 22kohm=22200ohms max.
Put these values in upper formula gives av varying between 2 and 111 or in dB roundabout 6 to 41 (dB=log10(av)*20). This 35dB difference divided by (20-1) steps giving 1.838 dB per step equaly divided.
Assuming leftmost stepposition is lowest amplification,
Step 1, av=2=6.00dB, R=22000 ohms
Step 2, av=2.46=7.82dB, R=14866
Step 3, av=3.04=9.66dB, R=10584
Step 4, av=3.76=11.5dB, R=7781
Step 5, av=4.64=13.33dB, R=5840
Step 6, av=5.74=15.17dB, R=4445
Step 7, av=7.09=17.01dB, R=3413
Step 8, av=8.76=18.85dB, R=2635
Step 9, av=10.82=20.69dB, R=2039
Step 10, av=13.37=22.53dB, R=1578
Step 11, av=16.53=24.36dB, R=1217
Step 12, av=20.42=26.20dB, R=933
Step 13, av=25.24=28.04dB, R=708
Step 14, av=31.18=29.88dB, R=529
Step 15, av=38.53=31.72dB, R=386
Step 16, av=47.62=33.56dB, R=272
Step 17, av=58.84=35.39dB, R=180
Step 18, av=71.71=37.23dB, R=107
Step 19, av=89.85=39.07dB, R=48
Step 20, av=111.02=40.91dB, R=0 ohm
This is the variable part of gain. The rest of gain is depending on wiring and type of your in-/output transformers.
-Harpo
 
[quote author="Harpo"]this is not an attenuator switch. It is part of a noninverting op-amp...[/quote]

Of course, I do understand (generally) that we are changing the gain of the amp by manipulating the feedback. Thanks for making the distinction.

[quote author="Harpo"]with voltage gain av=1+(Rf / Rx), where Rf is your fixed feedback resistor R1=22kohm.
Rx varies from R2=200ohms to R2+your rev log value of 22kohm=22200ohms max.[/quote]

That makes it perfectly clear.

[quote author="Harpo"]Put these values in upper formula gives av varying between 2 and 111 or in dB roundabout 6 to 41 (dB=log10(av)*20). This 35dB difference divided by (20-1) steps giving 1.838 dB per step equaly divided.
Assuming leftmost stepposition is lowest amplification,
Step 1, av=2=6.00dB, R=22000 ohms
Step 2, av=2.46=7.82dB, R=14866
Step 3, av=3.04=9.66dB, R=10584
Step 4, av=3.76=11.5dB, R=7781
Step 5, av=4.64=13.33dB, R=5840
Step 6, av=5.74=15.17dB, R=4445
Step 7, av=7.09=17.01dB, R=3413
Step 8, av=8.76=18.85dB, R=2635
Step 9, av=10.82=20.69dB, R=2039
Step 10, av=13.37=22.53dB, R=1578
Step 11, av=16.53=24.36dB, R=1217
Step 12, av=20.42=26.20dB, R=933
Step 13, av=25.24=28.04dB, R=708
Step 14, av=31.18=29.88dB, R=529
Step 15, av=38.53=31.72dB, R=386
Step 16, av=47.62=33.56dB, R=272
Step 17, av=58.84=35.39dB, R=180
Step 18, av=71.71=37.23dB, R=107
Step 19, av=89.85=39.07dB, R=48
Step 20, av=111.02=40.91dB, R=0 ohm[/quote]

Yes.... that looks much more like what I expected to see. Thanks for the clear explanation, and thanks also for running the numbers for me.

[quote author="Harpo"]This is the variable part of gain. The rest of gain is depending on wiring and type of your in-/output transformers.
-Harpo[/quote]

Got it. This information is a big help. Thank you again for your assistance.
 
Hi everybody !

I've got an idea for whose have problem when phatom power is shut-down ( when there is still 1 or 2 volts after stopping phantom power I mean).I've the same problem on my API and I was wondering to use a Zener diode, connected on the 48V path, such this one : http://fr.farnell.com/jsp/search/browse.jsp;jsessionid=0PTI0VHUIE0IBQFIAEXJLTQ?N=1000048+131830+108913&Ntk=gensearch_004&Ntt=diode+zener&Ntx=&_requestid=106069

This could let the 48V going to the mic input when running, and when the phantom power is shutted-down, the tension will descrease, and when it will be under 47V ( for the diode shown on the link below), the diode will stop everything .... what do you think about it ? does the diode will bear the current during running with 48V ?
 
I have a small question regarding the relais connectors/feet. I've seen on other BOMs that a Panasonic SMD model is recommended, is it possible to safely bend their connectors downwards, so I can fit them in the pcb holes?

thanks!
Christoph
 
Hi Crisotop

I think you can do it if you take care to not bend it to much.. i've already done it and it the machine still work (.. but maybe some people could tell us if it is a good thing or not ...?)

I have 2 questions :

I have to much "hum" and "buzz" noise in my API ( Fabio boards and PSU) at low gain .. and lots at high gain. this happens even if all connectors ground are connected in "star ground" to the PSU, I've put some copper shiled around each output trafo and around PSU trafo too. I use twisted cable for all audio connexions, and even when the box is closed I still have this noise ! Moreover I have a parasite sounds at high frequencies ( not oscillation, even if I have it at high gain with DI input :evil: ) ... and sme people seems to avoid it by changing regulators voltage (I use 24V relay).

Do you have an idea or some advices to avoid or to lower all this hum and buzz noise ? ( I don't want to put the PSU into an external box for the moment)

.. Does somebody found a solution to avoid oscillation of the 2520 op amp with the DI input ?? I've tried to change the 120pF caps by a 220pF but it didn't change anythig .. just seems to add some gain ...

thank for your help !
 
Use shielded wire for all audio connections and keep some distance to the PSU. What power trafo do you use?

Also, i think the input trafo and the OP's are most sensitive. Correct me if i'm wrong but i don't think copper will help against electro-magnetic noise since it's not magnetic.
 
Hi Flitan !


yes I think I'll change the cables for shielded ones. The trafo is toroïd ( but I don't remeber its characteristics) and I tried to put it as far as possible of the audio trafo and connectors. The copper sheet are used to make a "Faraday enclosure" .. but I didn't think it works well in my case because it needs to have electrical continuity all along the sheet ( a closed circuits, to allow eddy current. Do you think I need to have a shield on the Op-Amp case ?
 
Hi everybody !


I made some modification to the API but I still have this fu*** hum in the last quarter of turn of pot !

Here some pic to show you how look the preamp.

here a general view :

DSCN2841mod.JPG


As you can see, I use shielded audio cable only for DI input, pots wiring is made with audio cable too, but as Fabio told me, it needn't to be shielded.

The star ground is made as shown on the photo, and the Mic input XLR chassis are connected to the earth.


here more detailled pics :

DSCN2840.JPG


DSCN2838.JPG


DSCN2836.JPG


I still have "hum and buzz" in the audio path, so I'm wondering to add a metal sheet between the PSU board+ trafo and the rest of the audio circuit. Do I have made something wrong in the wiring ?

thank for your help and your advices ! :wink:
 
try moving the PT out of the box!
The AC inlet you used might insert hum tried them with out it.
Output transformers fit on the board if you take out the chassis!
 
Is one channel worse than the other? If so, which one? I posted pics of my dual Fabio 312 build in this thread (I think--of not check the gallery meta). I think I used shielded cable from the rotary switches (instead of pots) to the board. No hum problems on mine. You can probably trace my grounding in the photos as well. I believe there are quite a few photo sets like this around here, so check them all...

A P
 
[quote author="peter purpose"]Get that copper tape off the mains transformer toute suite..!![/quote]

Yeah, isn't that Dangerous?
 
Hi everybody

thank for the advices. For the Output trafo, I didn't saw it could fit into the PCB hole, so I'll try to move them into it. For the copper sheet on the PSU trafo it doesn't seems to be dangerous.. there is many layers of plastic on it, just under the copper sheet. I added it so as to make a kind of magnetic shield...but it doesn't seems to be efficient at all -lol.

I have the same buzz and hum on both channel (this is what suprise me a bit) but I have just little more high frequencies noise on the closest channel of the PSU board (which seems to be normal so).

I'll try some mods, thanks to you guys :grin:
 
From our very own ssltech....

"-just a quick note and as a very coarse rule of thumb... 50Hz hum (it IS 50Hz in your locality, isn't it???) usually means too many grounds, 50Hz Buzz means "too few" grounds (i.e. none!) 100Hz hum means inadequate ripple smoothing... If you're in a 60Hz zone, replace '50' with '60', also replace '100' with '120'.

You might have a power supply ripple problem... do you have a scope at all? -I've experienced Ripple due to inadequate reservoir cap size, excessive current draw, or -in odd cases- excessive power source impedance...

Keith"
 
Does the hum/buzz change when you switch phantom, polarity, mic/line? Are you using 5V relays? Does your relay rail regulator get really hot? Have you read all the stuff in the beginning of this thread about Fabio's PSU and the relay supply rail?

A P
 
Try to get rid of your bunsh of ground loops 1st. When you have a look at the trace side of the supply- and preamp-pcb's, you'll notice, that nearly all gnd's are already joined together (except relay-gnd on preamp-pcb).
I'd only use 1 wire for gnd from your power supply board to your star ground,
1 wire for gnd for +/-15V-gnd and phantom-gnd from each preamp board to the star ground
1 wire for relay-gnd from each preamp board to the star ground
YMMV.

Don't trust your solder joint of safety-ground to the bottom of your case. It will break sooner or later. Bolt a screw with solder lug and sawtoth washer instead and scrap off the coating of the case in this spot. This spot may also be dedicated as your one and only star ground, so yours don't have to fly.

From your pic you seem to use mono potentiometers, so a simple switch for stereo tracking will not work. You'll need at least a stereo pot in 1 chanal, else you'll nearly shorten both inverting inputs of your op-amp.

-Harpo
 
Hi everybody and thank a lot for all !

I answer to your questions in disorder :

->The pots are not mono but stereo, and the stereo link works very well :razz:

->I don't have scope here for the moment but It seems at my ears that there is little hum mixed with buzz .. this is strange. I don't have "big sub bass hum", but there are some, and the buzz sound is here. I have to mention that this noise become high only at high gain ( in fact in the last quarter of turn of the pot)...but I'll record this noise to make you hear, this would be easier than words lol.

->the noise is lower when I switch on Mic input., and get lower again if I put the -10dB pad ! ( this means that noise come before going trough pad, phase relay ?)

-> I use 24V relay and no 5V.

Firstly, I had to say that this API is all from Fabio (PCB, DOA) and I used his BOM too. Many guys have made the same project with me (we made a group order) and some of them haven't any problems at all. But one get the same buzz and hum problem, and used some copper sheet around the PSU trafo, which removed all noise !

About the power supply ripple problem.. do you think it can happen even if I use 2200UF caps (Rubicon) ? When I shutt down the power, the power LED still ligh during 20/30 seconds, I thought it showed that caps where very large and then it could be enough to filter correctly the current.. :?

I'd only use 1 wire for gnd from your power supply board to your star ground,
1 wire for gnd for +/-15V-gnd and phantom-gnd from each preamp board to the star ground
1 wire for relay-gnd from each preamp board to the star ground
YMMV.
Click to expand...

... I don't know if I understand right what you explain.. but in the actual wiring, preamp board ground is connected with PSU ground and star ground bunch... I'm sorry but I don't undertand really what is the difference between what you explain and the actual ground wiring :sad:

Should I connect all the grounds together (Relay, phantom power, +/-18V, and audio connectors) and then connect this bunch to a unique point on the PSU board ? (which means that I would remove all the ground cable between PSU board and each preamp board like is actually wired)

So, I'll try first to move the output trafo on the preamp PCB, and try to move the AC inlet. I wait to understand a little more this star ground stuff to make some mods :oops:

Thank to everybody for your help :razz:
 

Similar threads

frob
312 collected info, is this right?
Replies
1
Views
1K
soundguy
S
T
Calrec 1061 Filter Simulation - help needed.
Replies
7
Views
6K
TomWaterman
T

Latest posts

Back
Top