Poorman's Enclosure

[tomas1808]

After failing miserably to achieve acceptable noise levels with my DIY preamps (UA 1108, Green Pre) I figured that the cheapo enclosures I have been using are simply not good enough. Both preamps use the same cheap enclosure and both have exactly the same hiss noise levels. Since both have exactly the same amount of noise I think this is probably not a grounding problem but rather poor RFI shielding properties from the enclosure itself.

Since then I decided to redesign my projects, focusing exclusively on noise levels. First step would be selecting a suitable enclosure:

Aluminum 19" racks seem to be the most popular option but sadly this are impossible to get here in Uruguay, and importing these from abroad is prohibitively expensive ($150 for a standard 1U).

Looking for alternatives I found these extruded anodized aluminum enclosures: http://www.ebay.com/itm/230870547137?ssPageName=STRK:MEWNX:IT&_trksid=p3984.m1439.l2649

They seem extremely sturdy, probably have excellent continuity, made of aluminum (RFI shielding) and very affordable 



I figured I could use one of these for every future pre, comp, etc. I see them as PSU-less modules ala API500 but with their own enclosure.

All these modules would be powered by a single universal PSU (with ±16v, 24v and 48v like GDIY). This would save me the hassle of building a PSU per project and also help avoid the problems of onboard PSU (EM noise from power transformer).

The new VPR Floor Box PSU seems to be the way to go (albeit with a different case, I plan to use 5-Pin XLR connectors)



The purpose of this post is to ask if this is as good as I think it is.

Since I am a noob I figured there probably is something extremely wrong with this kind of setup...

What is the catch?

Is this a good idea?

Thanks!

 

[shabtek]

a central power supply is a good idea.

what are your current enclosures made of; could they be lined or retrofitted with sheet metal

i like to purchase blank front panels and fabricate or recycle the rest of the case

 

[tomas1808]

what are your current enclosures made of; could they be lined or retrofitted with sheet metal

I dont know what these are made of.

I guess I could retrofit them, but I rather replace the enclosure altogether.

I like the idea of having one standard "proven-quality" enclosure for all my projects, this should simplify the building process, so I dont mind replacing these.

I am searching for that proven enclosure. Hopefully the one I linked before is "the one".

If they were available here I would obviously go for rack enclosures and save all the hassle, but that is not the case.

i like to purchase blank front panels and fabricate or recycle the rest of the case

On a semi-related note, one thing that I like about these extruded enclosures is that I can attach up to 3 of them to a 2U front panel like in the pic. This gives rack compatibility for the future (when I hopefully have money to get racks)  ;D

BTW, how do you fabricate the rest of the case? Do you have sheet metal benders and other tools?

 

[ruffrecords]

I think these enclosures are made by a company called BOX and are available in a variety of sizes and colours from Farnell:

http://uk.farnell.com/box-enclosures/b2-080rd/enclosure-instrument-aluminum-red/dp/9307184

I recently purchased one for a tube mic pure project I am considering. They are also used as single units or in pairs in a rack mount panel in commercial products like these:

http://www.phaedrus-audio.com/philter.htm

I think they are an excellent range of good looking cases.

Cheers

Iaan

 

[tomas1808]

Iaan, thanks for the links.

Good to know that these are being used for pro audio.

I already ordered the ebay enclosure.

Lets see how it turns out.

 

[abbey road d enfer]

After failing miserably to achieve acceptable noise levels with my DIY preamps (UA 1108, Green Pre) I figured that the cheapo enclosures I have been using are simply not good enough. Both preamps use the same cheap enclosure and both have exactly the same hiss noise levels. Since both have exactly the same amount of noise I think this is probably not a grounding problem but rather poor RFI shielding properties from the enclosure itself.

Since then I decided to redesign my projects, focusing exclusively on noise levels. First step would be selecting a suitable enclosure:

Wrong! The first step is finding what causes this excessive noise. Poor EMI/RFI protection most often results in spurious or continuous noise, not white noise. What type of noise are you experiencing?
These "cheapo enclosures", are they non-conductive? Most RFI/EMI can be quite effectively shielded with simple aluminium or copper foil. OTOH, magnetic induction can be shielded only with ferrous material, but is your problem hiss or hum?
Have you measured the noise? Are you sure you evaluated noise with the proper input termination?

 

[mich]

Hi ,
Seems to me you made a good choise .the shared PSU is defenitly a great idea.solid state preamps and comps doesnt use a lot of current
anyhow (100mA Each tops?) and usually its +/- 15V and 24V and phantom.
so if you build lets say a +/- 18V@500mA, 24V@500mA ,48V@150mA - you will be covered for most of the SS projects here....
im just throwing numbers here ,but you get my point....
about your existing noisy boxes - probably you can hook them up to your Master PSU and be done with the noise too...(if thats the problem!)
with tubes its a bit more complicated cause when you start to add up filamets current - lets say 4 x two bottle preamp - you are already @ 2.5 -3 Amps ,and than you have B+...but i guess also can be done (and beeing done in consoles or the like all the time)

 

[tomas1808]

abbey road d enfer:

Problem is that I have run out of ideas regarding the cause of the noise! That's why I am starting to remove variables.

What I am getting is Hiss/white noise and sometimes I can hear very quiet radio stations (mostly on green pre). No magnetic induction problems.

I terminated the input of the my DIY UA 1108 with a 150ohm resistor between pins 2 and 3 of the XLR. The output is plugged into my converter (Audiofire12) and I am using Reapers spectrogram and meters. With gain at max the noise floor reads around -68dBFS which I believe translates to -50dBu (I think +4dBu is calibrated to -14dBFS). The gain of this pre is about 45dB. I am not 100% sure this is the right approach.

Regarding my current enclosure, I dont know what it is made of. The panels are all thickly power coated so I had to remove the paint in the contact points, according to my DMM there is continuity but I dont quite trust my DMM. Also the case doesn't feel very sturdy. In other words, it doesn't feel trustworthy at all.   

I know it is probably unnecessary to replace it but it will minimize the variables and also get a much better quality enclosure along the way.

Current enclosure pic attached.

mich:

Yeah I will be building solid state equipment exclusively, a single VPR PSU should last for a long time. Dont worry I wont be doing any tube stuff anytime soon!

The only thing I am not too sure about this approach is the PSU cabling. I dont like the idea of having 20+ power cables going from the PSU to their respective modules... quite messy I think.

Maybe I'll daisy chain the modules? Not too sure how to do that.

 

[abbey road d enfer]

What I am getting is Hiss/white noise

This is due to the input circuitry, could be a faulty tube, problematic layout, ...not the case.

and sometimes I can hear very quiet radio stations

This could be a problem with the case continuity

(mostly on green pre).

Solid-state is generally more sensitive to RFI than tube.

I terminated the input of the my DIY UA 1108 with a 150ohm resistor between pins 2 and 3 of the XLR. The output is plugged into my converter (Audiofire12) and I am using Reapers spectrogram and meters. With gain at max the noise floor reads around -68dBFS which I believe translates to -50dBu (I think +4dBu is calibrated to -14dBFS). The gain of this pre is about 45dB. I am not 100% sure this is the right approach.

This is the correct approach, computes at -95 dBu EIN, which is not correct. Do you have another preamp that you can compare against? Could be an interfacing problem between the mic pre and the interface...

Regarding my current enclosure, I dont know what it is made of. The panels are all thickly power coated so I had to remove the paint in the contact points, according to my DMM there is continuity but I dont quite trust my DMM. Also the case doesn't feel very sturdy.

Sturdiness doesn't matter. Only continuity matters for noise. Aluminium foil, flimsy as it is, is excellent for shielding.

In other words, it doesn't feel trustworthy at all.

Trust is for believers, not DIY'ers.   

 

[tomas1808]

Sadly I only have my DIY 1108 pre. I made the Green Pre for a friend, so its not at home. BTW the 1108 is a Class A FET pre, not tube.

In other words, I dont have anything to compare to.

I know sturdiness is in no way related to noise, but its good thing.

I want to believe 



PS: Curiously both 1108 and Green Pre had the same amount of noise at max gain. But I just remembered that the Green is supposed to have higher gain.. so.. 

 

[Rob Flinn]

I have used out dated computer equipment  or military surplus rack cases in the past.  Things like ethernet switches are often 1U cases.  I just replaced the front panel with a new one.  With the right boxes this is pretty easy to do, & once you bolt it in a rack nobody knows the difference.    Often the hole is already there for the iec socket too.

 

[tomas1808]

Interesting. I'll keep an eye.

 

[abbey road d enfer]

BTW the 1108 is a Class A FET pre, not tube.

Sorry, my mistake. Anyway, both FET's and tubes are less susceptible to RFI than BJT's.

PS: Curiously both 1108 and Green Pre had the same amount of noise at max gain. But I just remembered that the Green is supposed to have higher gain.. so.. 

That's what makes me think that maybe you have an interfacing problem. Could you make a more accurate assessment of gain. The actual gain of the 1108 depends so much on the xfmrs.

 

[ruffrecords]

abbey road d enfer:

Problem is that I have run out of ideas regarding the cause of the noise! That's why I am starting to remove variables.

What I am getting is Hiss/white noise and sometimes I can hear very quiet radio stations (mostly on green pre). No magnetic induction problems.

I terminated the input of the my DIY UA 1108 with a 150ohm resistor between pins 2 and 3 of the XLR. The output is plugged into my converter (Audiofire12) and I am using Reapers spectrogram and meters. With gain at max the noise floor reads around -68dBFS which I believe translates to -50dBu (I think +4dBu is calibrated to -14dBFS). The gain of this pre is about 45dB. I am not 100% sure this is the right approach.

Regarding my current enclosure, I dont know what it is made of. The panels are all thickly power coated so I had to remove the paint in the contact points, according to my DMM there is continuity but I dont quite trust my DMM. Also the case doesn't feel very sturdy. In other words, it doesn't feel trustworthy at all.   

Looking at your 1108 enclosure pic and the 1108 schematic:

1. The spec for the 1108 says its input noise should be -124dBm which at 45dB gain means the noise at the output should be about 45dB higher or around -79dBu. If you genuinely are only getting -50dBu then there is something seriously wrong.

2. I notice you appear to have a wire link from the pin1 of each XLR to the next and then to the ground star point. Pin1 of XLRs should always be connected directly to the chassis at the XLR.

3. The screened cable from the input XLR should have its screen connected to the XLR pin1 but be unconnected at the input transformer end.

4. The input stage of the 1108 is a FET. These are often very quiet devices but you sometimes get rather noisy ones. It might be worth replacing the FET.

5. The 1108 manual says it needs a really quiet power supply. It says a poor power supply can often add more noise to the output than the amplifier itself. Looking at the circuit it is not hard to see why. I would suggest you try adding a 1000uF/50V capacitor right across the power supply where it enters the 1108 PCB.

Cheers

Ian

 

[tomas1808]

That's what makes me think that maybe you have an interfacing problem. Could you make a more accurate assessment of gain. The actual gain of the 1108 depends so much on the xfmrs.

I am not too sure I know how to measure gain, or if I have the right tools.

Maybe I could send a signal from the Audiofire to the 1108, then back, and compare the difference in Reaper? Is that good enough?

2. I notice you appear to have a wire link from the pin1 of each XLR to the next and then to the ground star point. Pin1 of XLRs should always be connected directly to the chassis at the XLR.

Thats good to know. I did try a thick wire bus connected to both XLR with a cable going from the bus to the chassis. Maybe not good enough. I will try what you say.

3. The screened cable from the input XLR should have its screen connected to the XLR pin1 but be unconnected at the input transformer end.

Yes, the cable screen is not connected to the transformer, I left that end disconnected.

4. The input stage of the 1108 is a FET. These are often very quiet devices but you sometimes get rather noisy ones. It might be worth replacing the FET.

Yes, I ended putting a socket for the FET and tried about 10 J309s without any significant variation. I pretty much replaced most components on the PCB (and it shows, poor board :-\). I may get another board and start over again.

5. The 1108 manual says it needs a really quiet power supply. It says a poor power supply can often add more noise to the output than the amplifier itself. Looking at the circuit it is not hard to see why. I would suggest you try adding a 1000uF/50V capacitor right across the power supply where it enters the 1108 PCB.

Hmm, I built Mnats PSU which he designed exclusively for the 1108. I will try that though. Not too sure about what you mean with "across the power supply",  across the +24 rail and ground?

Thanks!

 

[ruffrecords]

Hmm, I built Mnats PSU which he designed exclusively for the 1108. I will try that though. Not too sure about what you mean with "across the power supply",  across the +24 rail and ground?

Thanks!

Yes, that's exactly what I mean but you need to place the capacitor close to the 1108 PCB rather than at the power supply PCB.

I noticed you said about your noise measurement that you made it at 'maximum gain'. Is there some means of changing the gain?


Cheers

Ian

 

[abbey road d enfer]

That's what makes me think that maybe you have an interfacing problem. Could you make a more accurate assessment of gain. The actual gain of the 1108 depends so much on the xfmrs.

I am not too sure I know how to measure gain, or if I have the right tools.
Maybe I could send a signal from the Audiofire to the 1108, then back, and compare the difference in Reaper? Is that good enough?

You have to find a way to make proper measurements. You could use a freeware such as rmaa or visual analyser by sillanum soft. Many are available. Find a way to produce a steady reliable tone at about -40dBu (you need some ingenuity there, by calibrating the output at about +10 dBu, using a standard bench or portable multimeter, then decreasing the level using the software's virtual attenuators). Then inject this signal into your preamp (there may be a few dB of attenuation due to the mic input loading the soundcard's output, but it should be less than 1 dB). You can then measure the preamp's output level with your multimeter. Caveat: some analog multimeters don't react well to high frequencies. Make sure that you use a midrange tone (250 to 400 Hz).

2. I notice you appear to have a wire link from the pin1 of each XLR to the next and then to the ground star point. Pin1 of XLRs should always be connected directly to the chassis at the XLR.

Thats good to know. I did try a thick wire bus connected to both XLR with a cable going from the bus to the chassis. Maybe not good enough. I will try what you say. 

What Ian says is unquestionable, but I think it is not the culprit. Noise due to RFI/EMI is either tonal - meaning you can identify discrete frequencies or transient - clicks, pops, or even plainly recognizable broadcast program! Continuous white-ish noise is intrinsic to the nature and topology of the circuit.

 

[tomas1808]

Ok.

So my objective is to try to achieve the -79dBu nosie floor @at max gain I am supposed to get with this circuit (provided it does have the expected 45dB gain)

I will try to measure the gain of my pre according to abbey roads d enfer instructions.

Also I will fix the XLR connections and try the 1000uF cap across the rails.

I noticed you said about your noise measurement that you made it at 'maximum gain'. Is there some means of changing the gain?

Yes there are currently two: The gain pot (controls a vactrol which LDR acts as a voltage divider) and a onboard switch that decreases the gain by 5dB. One curious thing about this is that it introduces a loud (inaudible on my speakers) 25Hz oscillation and its harmonics when switched to the lower gain setting. This is yet another reason why I am thinking of starting over with a new PCB.  :-[

 

[mich]

did you try short the input to ground?
if quiet when shorted - at least you know that you have no PSU problems.

 

[ruffrecords]

did you try short the input to ground?
if quiet when shorted - at least you know that you have no PSU problems.

Good point. I have an XLR with pins 2 and 3 shorted together for just this purpose.


Cheers

Ian