MILA-1 decent build pictures

[ChrioN]

Looks nice. What exactly did you change in the design?

 

[buschfsu]

didn't notice an impedance switch in daves design how did you implement that? (which Rs?)

 

[Kingston]

Looks nice. What exactly did you change in the design?

Some recap,

completely different PSU configurations, input configuration changes and additions, different input and output transformer loading (datasheets) obviously and output stage feedback range expanded.

I won't bother with PSU details. It was an experiment that works, but today I would do it different. The input module is pretty standard stuff too. Feedback range is easy: remove R18, and R19 could be configured as:

1. 50k (20dB NFB) (+5dB on input stage)
2. 50k (20dB NFB)
3. 100k (15dB NFB)
4. 200k (10dB NFB)
5. 470k (5dB NFB)
6. disconnect (no-feedback)

or a 1meg linear pot with about 50k in series with it.

those dB values are rough estimations.

more gain on no-feedback, not clean of course. It sounds best on around 200k-470k to me, depending on how loud the output stage is hit

NYD reports different gain values for the stages in his schematic. Mine measures a bit different, and hence the different values on the front panel (and above estimations).

didn't notice an impedance switch in daves design how did you implement that? (which Rs?)

Cinemag CMMI-10C datasheet. It's an additional tap on the transformer primary. Not that useful to be honest. Switch is DPDT relay.

 

[mitsos]

that's amazing man.... very nice build...

 

[andrew_k]

Wonderful!
Very encouraging, as I have a part-completed dual MILA build using the layout which you kindly shared. Add to that the fact I'm already using the same Russian PIO and that the PSU in mine is the G9 PSU with an LC filter following and I'm getting very excited that I will hopefully be as happy with the end result as you are!

 

[Kingston]

using the layout which you kindly shared.

I didn't post the layout here, because I did some changes to it afterwards. One of the most important changes is that the ground bus and B+ is now cut in the middle (used to be perfect mirror, tied in the middle), and both channels are taken separately to their own RC output of the PSU. Better channels separation, probably unnecessarily so.

 

[Holger]

Great build, Kingston. And yes, it looks better without the handles.

 

[MartyMart]

Gorgeous build, you should be proud of that one

MM.

 

[mrclunk]

Nice man, panel reminds me of Grace stuff.

 

[snipsnip]

very very nice.

Jake

 

[eskimo]

looks great. 

Great smudge/softening effect on drum bus as well and needless to say it works best on bass. I'd say this type of distortion is the main reason we still use tubes. Surprising how few serious preamps feature this. I mean, it's damn easy to drive tubes to distortion, and it seems most preamp designs do everything there is to avoid it.

So true.  Virtually nothing you can buy off the shelf with low NFB.  I think the Manley preamps are 10 dB NFB at the lowest setting; much lower than most.   25+ dB NFB was the norm in pro gear by 1955.  RCA BC-2B console pre was low for it's time, at about 10 dB.    My 95 dB Gates units of 1947 have at least 30 dB NFB.   1955 on you get units that clip harshly, not quite as harsh as the next decades transistor gear, but not too far off.   All depends on if you want euphonic drive soft knee, or not, as to which type you like. 

I think NYDave was paying attention to all the criticism out there about how the tone changes with NFB change, and limiting the available steps the way Manley and some others do. 

Sounds like there's a market niche to be filled.
Nice build Kingston!

 

[emrr]

There are people filling it, just not well funded people with advertising dollars.  and most are in the hi-fi realm. 

 

[eskimo]

Are there any studio examples?

I think a consumer unit would sell loads.
Say $500ish, in a GT brick size enclosure.

Weird that no big players are doing that. I suppose they don't dare to put those specs on paper?

 

[maxwall]

Nice work,

What caps ( ex: C05, C06 ...)  were replaced with the PIO's (audio coupling ?) and did you stay with same values ?

Were the PIO's used as replacements for basic mylars or orange drops in the original MILA layout ?

 

[emrr]

For one, the American Lundahl distributor had a DIY RCA BC-2B preamp kit, and discontinued it.  I never heard of many people buying it.  Looks like he's promising something new soon.  Lots of these hi-fi type dealers have kits that are easily adapted, I think. 

http://www.kandkaudio.com/

 

[mitsos]

Nice work,

What caps ( ex: C05, C06 ...)  were replaced with the PIO's (audio coupling ?) and did you stay with same values ?

Were the PIO's used as replacements for basic mylars or orange drops in the original MILA layout ?

looks like C3 and C4 according to his old layout diagram... C6 and C7 are also in the signal path right?

I suppose going a bit higher in cap value would be ok?  Would/should it make a difference?

I'm also curious about your PSU ideas.. almost have all the parts to build me a channel...

cheers! and once again, congrats! awesome build!

 

[Kingston]

I originally only had a misc selection of polyprop caps.

Now C3 and C4 are indeed PIO. C4 is changed to much larger one because I didn't have a small PIO like that at hand. I think I put in 0.22uF which is 10X the original value. But it's not going to make a difference anywhere. It's just interstage DC removal, and there's nothing that could resonate there ie. generate a bass bump. C7 (feedback DC removal) will be PIO as soon as I get a smaller 1uF to fit in there. C6 is 4,7uF polyprop, and as much as I'd like a PIO cap there, a 12cm x 6cm cap simply won't fit it.

 

[maxwall]

C6 is a prime location for a 5uf non-polar oil filled motor run cap. I have not seen any one try it yet
but sounds like it could be a interesting experiment. A good ole' well sealed Mallory or CDE one with the nasty PCB
tar inside would do nicely. I would'nt rule out the newer ones like Dayton and others without PCB fillers inside too.

When I get this dual rectifier PP guitar amp prototype off my bench , then I want to finally experiment with a MILA using all the
cap combos to see what sounds best.

 

[Kingston]

Upgraded the project a bit.

Only PIO caps in signal path now, except cathode bypasses of course but they are not in the direct signal path anyway (but a bad cap there will be heard nevertheless). There was plenty of room for even the huge 4.7-10uF PIO's, just not on the turret board. I'll still have to fashion some physical support for them. I guess a plumbers supply store would have a nice selection of brackets for these "pipes". Tested some different types of Russian PIO's. These are all very cheap compared to modern polypropylenes, but even the lowest end "wax paper" PIO's sound better than any polyprop I have (and I have a decent selection). Quite easy to spot the difference when there's four caps in the signal path because the amp is so linear. Will post pictures later. Cold war technology rules, and I love the idea these are originally made for radars, missiles guidance systems etc. extreme usage.



But dammit, still was not perfectly happy with the result. Turned out the tubes had worn out their welcome, or something. I have a decent batch of NOS RCA 12AV7. All of them sound like they are dead on arrival, so maybe it's this particular tube brand that sounds bad, and has nothing to do with wear.

I put in Tungsol 12AV7 and Jan philips 5965A and it's like someone turned off a nasty RFI interference in the signal path. NOW I'm done. It finally sounds "like a record".

Those Jan philipses and tungsols look completely different on the inside compared to the RCA's. The RCA's are all "tiny" on the inside, and have half the physical support for the internals (it's a computer tube originally). One could surmise it's for these reasons the RCA's can't hold up, or maybe I just had bad luck with my batch of 12. And maybe there's more than one type of RCA 12AV7.

That's the biggest difference of the "same tube" between two brands I have ever seen. It's not really visible but the rca is not just short, but a very thin and weak structure compared to tungsol as well. For the decent collection of tubes I have, I can only observe minor adjustments between models. The factories usually seem to copy each other quite strictly.

Of course we can't just assume all RCA's behave the same, no matter how they look like. I have 12 and they all behave the same. Not of statistical significance, but something to think about nevertheless.

I guess the lesson of the day is, don't assume a certain tube brand model is good just because someone says it is (RCA has good reputation in general).

PS. with the above sales pitch I now have RCA 12AV7's for sale.  

cheap.

 

[Kingston]

and here's the final layout. "Plumbing brackets" for the huge PIO pipe bombs not yet connected.

Good shielding of the wires to and from the big caps ensured there was absolutely no change in noise compared to the original positions of the small polyprops within the turret area. Noise floor is still black as a black mans cape.