RCA Ba2A Pre amp qestions & answers

[letterbeacon]

Looks like I edited my post at the same time as you were posting yours!

Thanks for the compliments on the build - I'll get some proper pictures up once I've finished the front panel.

It's supposed to in theory.  But bigger Q:  Do you have CT connected straight to ground?  If so, try installing two 100r 2W resistors and leave the CT tied off, or better yet, install a 100-200r pot.

Do you mean connect the two 100r resistors between the 6.3VAC and ground?

Before attempting I would go over the layout and make sure the filament wires aren't too near any sensitive audio wires.  I've personally never had filament hum due to layout issues and have even tried to purposefully induce it by pushing fil wires into close proximity of audio wires.

There is one instance (from V2 to the output TX) where the signal wire is fairly close to the heaters, but I used shielded cable at that point, to be on the safe side.

I'm going in reverse order here - rule out tubes by swapping, etc.  Also make sure that tubes are well seated into their sockets.  I just had a case of this on the recent preamp I built.  Brand new sockets that felt bitch tight, but tubes weren't all the way pushed in.  It was creating hum in output and a simple reseat cured it.

I'll try reseating the tubes, but in the back of my mind I'm thinking about the missing shield pin on both my 1620s!

Have you done step wise troubleshooting to isolate what stage hum is coming in?  I'm assuming maybe 1st since you say you have to crank it to hear it.  You can try to check output stage by shorting V2 grid to ground (or just turning vol pot all the way down), then run another small amp of some type on the output and bring vol up and see if it's still there in the same place.

So my plan will be to have the volume pot right down and then connect the output to one of the pre amps in my interface to and crank it up.  If there is hum this means it's coming from the 2nd stage, right?

 

[lassoharp]

Do you mean connect the two 100r resistors between the 6.3VAC and ground?

yes

There is one instance (from V2 to the output TX) where the signal wire is fairly close to the heaters, but I used shielded cable at that point, to be on the safe side.

The quick test there is to use a chopstick and push the wires away from each other and  see if it changes hum level

I'll try reseating the tubes, but in the back of my mind I'm thinking about the missing shield pin on both my 1620s!

I forgot about that.  I suppose it's a possibility.

So my plan will be to have the volume pot right down and then connect the output to one of the pre amps in my interface to and crank it up.  If there is hum this means it's coming from the 2nd stage, right?

With this method it will be showing any residual hum in 2nd stage - there could still be hum originating upstream too.  You'd be looking for any significant amounts here. If you have virtually none down to a low low level -80db?  Then likely it's all originating from somewhere in first stage.  If equally distributed in 2nd stage then maybe pointing to heater or PT related hum.

Do you have a  rough db figure for the overall hum level when you have it cranked? 

 

[letterbeacon]

With this method it will be showing any residual hum in 2nd stage - there could still be hum originating upstream too.  You'd be looking for any significant amounts here. If you have virtually none down to a low low level -80db?  Then likely it's all originating from somewhere in first stage.  If equally distributed in 2nd stage then maybe pointing to heater or PT related hum.

With the volume pot completely off, there is no noise whatsoever when I crank the inputs of my MR816 right up until the gain on the interface is up fully.  Then there is a hiss but to me that sounds like the noise floor and nothing like the hum I'm hearing when the volume pot on the BA-2 is up.

Do you have a  rough db figure for the overall hum level when you have it cranked?

With the volume on the BA-2 fully cranked, looking at the meters in my DAW, I'm getting around -3dB.

When I took the frequency response measurements I didn't have the amp cranked so as not to overload the convertors on the interface.  Thinking about it now, the noise figures aren't an accurate figure on the frequency response if I don't have it cranked.  D'oh!

I'm going to try and head to the shop today to get some resistors to try the artificial CT and maybe elevating the heaters.

 

[DaveP]

Hi Letterbeacon,

Just noticed your post about no Pin 1 on the 1620.  I have almost 100 RCA metal tubes of all types and every single one is earthed through pin 1.  Look at the original RCA spec on Frank's site, it clearly shows pin 1 earthed.

To have an unearthed chunk of steel covering the most sensitive part of your circuit is very bad news.  I think you have been sold some tubes that someone has messed about with.

A 1620 is just a low microphonic version of a 6J7.

I buy all my tubes from http://www.vacuumtubesinc.com/ in florida and if you keep the total below $18 you dont pay any customs duty or VAT.

best
DaveP

 

[lassoharp]

With the volume on the BA-2 fully cranked, looking at the meters in my DAW, I'm getting around -3dB

If I'm interpreting that figure correctly that sounds crazy loud for hum?


Q: Is the input terminated when hum is being checked?  If not, clip on a 150 or 250r resistor to the pri or plug a microphone into it (cover capsule with cloth to keep out room noise) and see if that makes a difference.

 

[letterbeacon]

To have an unearthed chunk of steel covering the most sensitive part of your circuit is very bad news.  I think you have been sold some tubes that someone has messed about with.

I think you're probably right.  I ordered some more in preparation for my BA-6A build, so they should be arriving soonish.

With the volume on the BA-2 fully cranked, looking at the meters in my DAW, I'm getting around -3dB

If I'm interpreting that figure correctly that sounds crazy loud for hum?


Q: Is the input terminated when hum is being checked?  If not, clip on a 150 or 250r resistor to the pri or plug a microphone into it (cover capsule with cloth to keep out room noise) and see if that makes a difference.

Yes, that's -3dBFS when fully cranked.  The reason I didn't notice it being so high before is because I never fully cranked it before.

No that wasn't with the input terminated, I'll have a go at plugging a mic in when I get some time tonight.

As we have established the hum is coming from the first stage, does this mean that the hum is unlikely to be coming from the PT?  The input stage is on the far side of the case to the PT, with the second stage 1620 and output transformer in the middle.  If the hum was coming from the PT then surely hum would be present in the second stage and output tx as that is physically closer to the PT.  Is that correct?

Another thing I want to try is to use shielded cable from the input XLR to the input TX.  At the moment I'm using tightly twisted wires.

I'll also try and get a diagram of my grounding system up tonight, to see if you guys can spot any school boy errors.  Basically the signal ground and power ground are starred at the same point on the chassis.  Both XLR pin 1s are connected to the chassis and all the shields (the shielded cables, the shield tap on the input tx etc.) are connected to pin 1 of the output XLR.

 

[letterbeacon]

I had some time to work on the BA-2 tonight - I've got the new tubes in, I've put shielded cable between the input XLR and the input transformer and I've put 100r resistors in to create an artificial centre tap for the heaters.

I've now got the hum down to about -27dBFS when fully cranked with a microphone plugged in (with many cloths covering the mic).  So I'm pleased with that, however I'm sure I can get it lower.  What sort of figure should I be aiming for?

I was thinking of trying to elevate the heaters by about 50VDC by using a potential divider across a feed from the B+.  It's the first time I've ever done this, is there anything to watch out for when doing this?

 

[emrr]

I am missing something about your hum measurement.  It appears very good compared to system noise in your plots, unless I'm seeing them incorrectly.  There it appears nowhere near the level you are mentioning, which would be unusable.  What am I missing? 

You shouldn't use a mic at all when doing this test, just a 150-250 ohm resistor. 

I didn't think elevated heaters addressed hum, only other noises.

 

[letterbeacon]

Apologies, I'm not being very clear.

When I made the plots I didn't have the gain of the amp fully cranked.  I only had it turned up enough so it wouldn't overload the inputs to my interface.  Therefore the noise plots are unreliable because they only show the level of the noise when the gain knob is about a quarter of the way up.

When I later did some real world tests recording an acoustic guitar with a ribbon mic, I found the hum to be very loud especially when multitracking.

lassoharp asked me:
"Do you have a  rough db figure for the overall hum level when you have it cranked?"

So I tried it the other day and got -3dB.  Tonight after using new tubes, shielded cable between the input XLR and input TX and with the artificial centre tap I got -27dB.

I'm slightly loath to rotate the PT as I've (stupidly) fastened it so it's a pain to move.  I will try that as a last resort though.  The thing that makes me think that it isn't the PT is that the second stage, which is closest to the PT, has no hum in it when the gain pot is turned right down (off).

You shouldn't use a mic at all when doing this test, just a 150-250 ohm resistor. 

Gotcha, I'll give that a go tomorrow -work has a habit of getting in the way of my DIYing!

 

[emrr]

Gotcha. 

The hum is still coming from the power transformer, one way or the other.  If the volume is down or off, the input transformer (most likely to pick up hum of anything) is not part of the picture.  Nor is the gain of the first tube there to amplify anything of low level.  The 2nd tube and output transformer are themselves alone likely not accountable for any gain at all, more like unity gain.  Orientation of power transformer and input transformer relative to one another are key.  Gates used that transformer on a 75 dB preamp with built in PSU, and hum was acceptable in it (I heard several of them), so it can be done. 

 

[lassoharp]

So I tried it the other day and got -3dB.  Tonight after using new tubes, shielded cable between the input XLR and input TX and with the artificial centre tap I got -27dB.

That's quite a change.  So, that's 3 new things.  First thing, differentiate some of what's doing what.  Put in old tubes with new CT & shielding and note change. 

  To further sort out possible filament balance issues a balance pot will be most helpful.  I've only really had to use a hum pot once for a final build and I'm inclined to feel that the 2 100r resistors does adequate job in most cases.  If it's not too much trouble I would install a balance pot and see if it helps get the figure lower.  I've found that on PTs where a simple CT or 100r to ground does the job, that if a hum pot is used you will notice nearly no noticeable appearance of hum until pot is rotated to the extremes.  On the one circuit that was troublesome the hum level changed perceptibly and constant through the range and the null was within a smaller range.

If rotating iron I would try to rotate input instead of PT if possible though it looks to be fairly tight for both.  Maybe a perfect 90 deg rot will do the trick.

 

[DaveP]

Hi letterbeacon,

This is a straightforward way to measure hum.

Put a standard 200 ohm metal film resistor across the input terminals (inside an XLR plug if possible).

Put a DMM across the output load (620 ohms?) then turn it up fully and measure the level.

A V77 amp is about 4mv so that is the kind of amount to expect.

Next check the gain of the amp with a 1kHz signal.  Now divide the hum level by the gain to get the E.I.N. (equivalent input noise).  Say the gain is 60dB and your hum is 7.7mv, then your EIN will be 7.7uV, this is 100,000 times less than 0dB which is -100dB.  The V series amps are -110 to -120dB for comparison.
If its a cheap DMM then it won't measure high frequencies very well but it will sure measure 50~120Hz as mains frequencies were what they were designed for.

best
DaveP

 

[letterbeacon]

Thanks all three of you for your replies -it's fascinating to me this whole troubleshooting process.

Orientation of power transformer and input transformer relative to one another are key.

If rotating iron I would try to rotate input instead of PT if possible though it looks to be fairly tight for both.  Maybe a perfect 90 deg rot will do the trick.

I think I'm going to have to bite the bullet and try rotating the PT at some stage.  Unfortunately the input TX can't be rotated as it's too wide to stand vertical.

That's quite a change.  So, that's 3 new things.  First thing, differentiate some of what's doing what.  Put in old tubes with new CT & shielding and note change.

Good point, I should have done it in stages and noted the stages really.

Next check the gain of the amp with a 1kHz signal.  Now divide the hum level by the gain to get the E.I.N. (equivalent input noise).  Say the gain is 60dB and your hum is 7.7mv, then your EIN will be 7.7uV, this is 100,000 times less than 0dB which is -100dB.  The V series amps are -110 to -120dB for comparison.
If its a cheap DMM then it won't measure high frequencies very well but it will sure measure 50~120Hz as mains frequencies were what they were designed for.

I don't quite understand the maths behind this I'm afraid.  You say divide the gain by the level of the hum.  So 60 / 7.7.  You get 7.7uv, but I get 0.128mv.  I'm obviously doing something wrong here, but I can't figure out!  Do you mean divide by 7.7mv expressed in dB?

Unfortunately I've got family commitments this weekend so I'm going to struggle to find time to do what you guys recommend, but I'll report back as soon as I do!

Thank you!

 

[DaveP]

Letterbeacon,

You simply mis-read my post,  for arguements sake say the gain is 60dB or 1000 times.

Next measure the hum/noise on full gain, say its 7.7mv.  You simply divide this by the gain 1000 to get 7.7uV,which was what it was before it was amplified.

I used 7.7 because the reference point 0dB is 0.77 V (actually 0.77459V) so it just makes the maths easier.

Obviously your figures will be different from this and at some stage you will have to convert a number to a log and multiply by 20 to get dB.

best
DaveP

 

[letterbeacon]

Thanks Dave, I understand that now.

Unfortunately I don't think my DMM goes low enough to measure mV accurately.  I tried the following but with not much success, I'll write out what I did in case I'm doing anything obviously wrong.

First I measured the gain of the pre amp by outputting a 1kHz sine from my DAW @ -80dBFS.  I ran it through the BA-2, with the gain up full, and back into my DAW.  The incoming sine measured -3dBFS meaning my BA-2 has a gain of 77dB.

Then I soldered a 200 ohm resistor across the + and - terminals of the input XLR and a 620 ohm resistor across the secondary of the output transformer (I couldn't easily reach the output XLR, but I'm pretty sure this does the same thing).  I then switched my DMM to it's 2VAC setting (the lowest it goes) and measured across the output tx secondary.  My DMM has three decimal places so it can in theory show mV, but I was getting 0.000VAC.

It's funny because when I had my leads not connected to anything, the DMM would read random numbers as i it was reading something in thin air (maybe the AC fields) but as soon as I connected the leads to the output tx secondary it would start to read 0.000VAC.

Have I done something glaringly obviously wrong here?  If not, it might be that I need a better DMM with a higher resolution.  I'm working at the BBC for a couple of days this week, I might see if I can use some of their equipment (if there's any left these days!).

So I tried it the other day and got -3dB.  Tonight after using new tubes, shielded cable between the input XLR and input TX and with the artificial centre tap I got -27dB.

That's quite a change.  So, that's 3 new things.  First thing, differentiate some of what's doing what.  Put in old tubes with new CT & shielding and note change. 

Now I'm using a 200 ohm resistor across the input I'm getting a reading of -33dBFS, I guess the 6dB difference must have been the room noise the mic was picking up.

I tried swapping out the tubes, the new CT and the shielding between the input XLR and the input TX and the tubes and the artificial CT don't do anything noticeable.  It is the shielding between the XLR and the input TX that is reducing the hum considerably.

I noticed that the wire coming from the first 1620 plate runs fairly close to the AC heaters wire.  I used a pencil to move the wire away from the heaters while the unit was on but I didn't notice anything change hum wise.  I'm thinking I might shield the wire anyway though, just in case.  It's only a about 3" long so hopefully capacitance won't be an issue.

Thanks again for sticking with me on this - it's been hugely educational so far.

 

[DaveP]

Hi Letterbeacon,

It wasn't a good day to go visiting was it!  What with practically blind drivers leading vast convoys of cars at 29mph and wind and rain to boot eurgh!

Anyway, its good news that you can't see any AC on your DMM, you would be in trouble if you could!

best
DaveP

 

[letterbeacon]

Hi Letterbeacon,

It wasn't a good day to go visiting was it!  What with practically blind drivers leading vast convoys of cars at 29mph and wind and rain to boot eurgh!

Ha!  No, it wasn't too pleasant!  More bad weather on the way down from Scotland apparently too!

Anyway, its good news that you can't see any AC on your DMM, you would be in trouble if you could!

best
DaveP

I think I may have misunderstood - wouldn't the hum be an AC (audio) signal? I thought I needed to measure the AC (in mV) present and then convert it to dB to figure out how loud the hum is in my pre amp.

 

[DaveP]

Letterbeacon,

No I mean that its good news that the hum is so low it doesn't show on your DMM!

best
DaveP

 

[letterbeacon]

Aha!  Sorry a bit slow here tonight!

 

[emrr]

so.....
.....why does your build have 27 dB more gain than it should?  Am I reading that correctly?