Valve / SS Hybrid EQ with External PSU – Checklist?

[thermionic]

Hi,

I’ve got an LC-based EQ on the bench which has a 24v discrete front-end and a valve output-stage with 280v HT supply and unregulated DC heaters.

It has an annoying 100Hz hum…. It’s ok with the music playing, but if 0dB is the peak, it would be around -60dB…

I’ve looked for ripple on the DC rails and they’re clean-as-a-whistle, which leads me to suspect a local ground loop…

Having an outboard PSU on a valve design is obviously more complex than keeping the PSU internal, so are there any obvious faults that you reckon might exist here? As a rule, I have always found solid-state designs to be quieter with external supplies, but for valve gear, I can’t think of too many precedents for outboard supplies…

The heaters are DC, via a raw unregulated supply (secondary – to - bridge rectifier – to - filter cap – that’s it).

I’ve tried powering the heaters from a bench supply and it makes no difference.

Obviously, the PSU case and EQ case are grounded. The head unit gets its chassis ground from the tag on a 5-pin XLR, which gets its ground from a tag on the PSU’s case also connected to mains inlet’s ground. I’ve taken resistance measurements between all chassis points and main GND – no issue there.

Any suggestions?


Many thanks in advance.

Justin

 

[NewYorkDave]

but for valve gear, I can’t think of too many precedents for outboard supplies…

Guess you haven't seen much old studio-grade valve gear :wink:

Outboard supplies were the norm for everything but the smallest systems (e.g., "compact" radio consoles).

My first suspect would be that raw DC heater supply but it seems you've eliminated that. However, you did confirm the heater supply is referenced to B- at some point, right?

Also, are chassis ground and B- tied together at one point only?

Did the unit ever work properly? Is it a commercial unit or a DIY project?

Didya try a new tube, just for laughs?

 

[thermionic]

>Guess you haven't seen much old studio-grade valve gear

Don’t dangle the bait, we need names! You’re talking about consoles, right? ADL make a Fairchild 670 clone with outboard supply for $30K USD – I guess I could buy one and see how he does it :grin: Manley make a Phono pre with o/b supply – a snip at $10K USD. The Telefunken consoles had external supplies – another cheap one to investigate!

>Outboard supplies were the norm for everything but the smallest systems (e.g., "compact" radio consoles).

Watch what you’re saying about my system!

>My first suspect would be that raw DC heater supply but it seems you've eliminated that.

Definitely.

>However, you did confirm the heater supply is referenced to B- at some point, right?

IIRC, yes, but I’ll check again tomorrow.

>Also, are chassis ground and B- tied together at one point only?

Again, I think I checked, but I’ll see tomorrow.

>Did the unit ever work properly? Is it a commercial unit or a DIY project?

It’s a prototype that I paid a professional designer to design. I built it as a 5-band EQ on strip-board (‘twas fun - not - took 3 days solid).

The designer is respected for his valve and SS units (he’s not unknown to the board here), but – being a successful designer – he can charge accordingly… I need the learning, so I'd rather have a bash myself before going back to the designer - tail-between-legs...It could really cost if he has to fix it.

>Didya try a new tube, just for laughs?

It’s stereo and both channels are the same.


Thanks :guinness:


Justin

BTW - If it works well, I *could* sell it as a commercial product (with a different supply for govt. reasons), so please don't give advice here if you're not cool with assisting on such a scheme (I'm not sure how different this is from asking about mic-ing a bass - you sell the music, right?).

 

[NewYorkDave]

Nah, man, I'm not talking about the nouvelle boutique stuff... I mean the real deal from the '50s or '60s, e.g. Langevin.

For a modest but functional and quiet example of a tube amp with an outboard supply, see my MILA-1 preamp:

http://www.twin-x.com/groupdiy/thumbnails.php?album=44

 

[thermionic]

Hi Dave,

Both sides of the heater supply are bypassed to B- via a low-value resistor on each leg.

The anode supply has an extra filter consisting of a 2.7K resistor in series with 22uF cap, going across B+ and B- where all grounds converge.

Chassis GND and B- only meet at the 5-pin connector on the head unit.

I just had a telephone chat with the designer and he’s suggested a couple of other things to try. I haven’t 'scoped the HT line yet (my scope only goes to 250v, so I’m making a divider to drop it down to look at it). He also said to try the HT transformer out of the case, away from the +24v line.

If this EQ made it to the commercial stage, I would pay the designer to make something elaborate with all manor of regulation. Because it’s just a prototype, I’m trying not to incur extra expense until I know the idea works. At the moment, I can’t take an EQ with only 60 usable dB into a studio for evaluation….

It should be noted that the designer installed the HT supply, after I built the main board. He didn’t spend time working on noise because he didn’t want to incur expenses until I had approved of the general sound. I think it sounds good, but I would rather avoid paying him to design a monster regulated supply until it gets a thumbs-up in the studio.

The current - albeit basic - supply we have is similar in scale to the one on MILA1, so whilst it may not be the world’s most high-end supply, it should surely be quiet enough to trot to a few studios…(MILA1's figures look v. impressive btw)

It’s been suggested that the surplus transformer (Danbury) we used could be a culprit as they’re cheap…

Anyway, I intend to spend a few more hours working (I need the experience). If I can’t fix it, it’ll be open-cheque book time again... If you want to develop professional products, you should expect it I guess… Having said that, I wired up all the PSU on the Mk1 prototype and there’s no hum at all. I’ve never had a hum beat me in the SS world in my life (I've made tens of projects in the SS world, and I've even assisted OEMs in making their SS designs quieter - can you believe that!), but I lack experience in building valve gear…

I have a terrible suspicion the cure could be very obvious...

BTW - MILA1 looks pretty cute. I reckon you could sell it as a kit. Considering how fashionable valves are these days, I think you’d sell a few. I can envisage small studios viewing the kit as an entry into the world of real valve gear (i.e. proper stuff with transformer coupling and HT plate supplies – not “marketing valves”).


Thanks again.

Justin

 

[thermionic]

FWIW, I have tried a few tests tonight to no avail...

I removed the HT transformer from the enclosure (heaters still wired from the bench PSU) and placed it a foot away, in case there was some kind of EMI thing going on, polluting the 24v supply near it (the waveform from the 24v rail is really clean, so this is out of paranoia really). No change whatsoever.

I put a 'scope probe on the 280v B+. There was a sinusoidal-type waveform, but it was the same whether B+ was switched on or not... It diminished if I turned off all the lights / iron etc in the workshop, so I think it's just 50Hz junk in the environment getting into my cabling.

The next (desperate) idea is to try isolating the output transformer brackets from the case and see if anything happens there (I told you it was desperate).

Next steps will be to visit a homeopathic electrician and try burning incense whilst sacrificing a sheep.

:roll:


Justin

 

[mikka]

Try another transformer..... a known good one ....just a thought.

 

[gyraf]

Classic problem with inductor-based eq's - this is why Klein & Hummel boast about their UE-100 being without inductors, and thus lower noise..

Get a very-low-noise toroid power transformer made. And then shield, shield, and shield.

Also, distance to transformer is your friend.

Jakob E.

 

[thermionic]

Classic problem with inductor-based eq's - this is why Klein & Hummel boast about their UE-100 being without inductors, and thus lower noise..

Even if the PSU is several feet away? The first prototype we made (I assembled it) was entirely discrete with a similar design of LC EQ (similar to this, but with a Neve o/p stage that I bodged in there for convenience instead of a valve affair) and boasted around 128 dB of range - with an internal supply...

Admittedly, this unit has a valve o/p stage, but the PSU is several feet away from the inductors...do the usual inductor-based noise issues apply in this scenario?

Thanks for the advice :guinness:

I think I'm going to try a decent toroid anyway.


Justin

 

[gyraf]

PSU is several feet away

yea, ok - - that might not be the problem then.. :razz:

The problem I mention usually only turns up in the very-large low-frequency inductors anyway..

I would check on a scope what sorta characteristic the noise has, and compare to the different supply voltages to find the supply line that it may come from..

 

[kevinc]

Another thought: Sometimes when there is ss stuff that is oscillating at 1MHz+ due to some kind of instability, the manifestation in the audio band is hum at the power line frequency at anywhere from a low level (-60dBish to as much as -30dBish). I've seen it a couple of times. Of course, I tortured myself about inductive coupling and ground loops first, before realizing what I was dealing with. :sad:

Probably a longshot, but worth a mention.

 

[thermionic]

Hi Guys,

Thanks to all for the help.

By pure chance, I did something last night that cured the hum… I knew the fault was something simple – the problem being, I ignored the obvious… the really obvious…

On the bench, I use a CD player and Denon amp for testing. They are both unbalanced… Both have 2-core mains cables, no earth…

The EQ is transformer I/O…

The cables to connect player and amp connect via hot + cold, leaving the shield floating at one end…

Can you see where this is going?

When I first got the hum, I was wondering why I couldn’t find it on my ‘scope (which has a 3-core cable…)… Doh! Double Doh!

That was until I connected a different cable, where the cold was shorted to ground…

About 18dB of hum rolled off immediately… According to my analysers, the EQ has a very healthy noise floor now…

In light of this “revelation”, I shall, of course, make a donation to the PFPH

(The PRR Fund for the Prevention of Hackdom)


Don’t call me if you suspect you have an elephant in the room, I’m the wrong guy.

It should be noted that the designer’s input has been nothing short of 100% professional and his ground scheme worked all along… I am the weakest link – goodbye!

BTW – I also gave myself a 300v “wake up call” finding this out… I forgot to drain the reservoir caps with a bleeder resistor (something I’ve been doing for 20 years, since I first built power amps) and got a major bolt from the PSU – it burnt my finger… My DVM measured 288v still lurking in the supply… If anything positive has come from this, it’s to remind the group that, if you don’t have bleeders as standard, keep a high-wattage resistor on a pair of croc clips in your pocket at all time and remember to drain, even if the cord is disconnected… Don’t get your fingers burnt…(literally).


Cheers!
Justin