• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

ready to give up

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
i'm coming up on my third month of trying to get the hum in my first tube kit down to a listenable level. the amp in question is an ella from diyhifisupply. i had a thread on their AA board, which can be found here.

i finally broke down and rebuilt the thing from the ground up, to no avail. i still have hum. confoundingly, it measures only 3.6mv and 4.8 mv in the L/R channels, respectively. i know that's pretty low, but i can still hear it from the listening position 2+ meters away. my speakers are only 90dB efficient.

i feel as if i've run out of things to try here. i have yet to swap out tubes, because this is my only tube amp and i have no spares lying around - i would have to order some. plus, i have hum in both channels, and i imagine the chances of having two bad tubes are pretty slim.

at this point, i'd appreciate any ideas/advice/offers to buy the damned thing :bawling:

-dave
 
Hi there,

I'm just guessing here as I'm not familiar with your amp but...

If I'm reading your post on audioasylum correctly, the hum got worse after you 'redid some of the grounding points'. Did you add to or change any of the grounding points while you were doing this? e.g. additional connections to the chassis, extra wiring etc. If so, you could be looking at a ground loop. Most valve amps need to be wired in a star ground arrangement - all the grounds return to a single point. If this has been altered you might get some noise problems as a result.

In a few of my own projects I had really bad hum because the heater supply was floating an not referenced to a similar voltage to the cathode (usually ground). This affects pre and driver stages the most and can vary with the tubes you use as well - very confusing
Check that the heater supply is referenced to something - this is normally done either by connecting one side to ground, a resistor network, or a capacitor depending on the design so check the schematic.
 
Member
Joined 2004
Paid Member
Tokyofist:

Voltsecond, a frequent contributor to the Bottlehead forum, maintains a website where he provides a number of suggestions for managing hum on the Bottlehead Foreplay preamp. You might try checking his website out -- perhaps there's something that could easily be implemented in your situation that would reduce your Ella's hum.

Regards,
Scott
 
Ex-Moderator
Joined 2004
Is the hum 60Hz or 120Hz? In other words, is it coming from the power tranny (primary or secondaries, including heater circuit) or from the B+ filtering? I know nothing about the Ella (except that it's got a silly name ;) ). However, the cures for 60Hz hum are not the same as those for 120Hz hum.

If it's 60Hz, check for unshielded signal-bearing components (e.g. decoupling caps, grid resistors, long grid wires) in the vicinity of unshielded or untwisted AC bearing leads. Are your OP tranny and power tranny magnetic fields at 90 degrees to one another, as they should be?

Is your heater circuit grounded at its centre tap? This can be achieved either by a centre tap on the 6.3 v winding, if there is one, or by using a pair of, say, 100 ohm resistors bridging the 6.3 v line with their junction grounded. A pot could be substituted for the 100 ohm resistors, with its slider grounded; this enables you to choose the position of least hum (in theory, anyway).

Have you tried connecting the heater centre tap to a voltage divider from B+ to ground, so that the heaters are at about +20v, with the centre tap grouded also via a 10uF capacitor? This prevents the tube heaters from emitting electrons and interfering with the operation of other electrodes.

Another problem that can cause 60Hz hum is if one side of your rectifier bridge or full-wave rectifier is not functioning.

If it's 120Hz, this suggests that your B+ filtering is deficient. You could try changing, or increasing the size of, the smoothing/filtering capacitors and/or using chokes in place of resistors in the filtering line.

A very common cause of hum, which can be difficult to diagnose, is a poor grounding scheme. Hum can be picked up easily by the sensitive early stages of an amp if there are ground loops or bad grounding connections. Using the chassis as ground at more than one point is a very bad idea. Multi-point grounding is the best approach, whereby all ground connections for each individual stage are to the same point; the individual stage grounding points are then connected together. The idea of this is to ensure that ground currents do not flow between different stages. Someone else can probably explain it better!

Good luck.
 
You need a triple star ground for max hum reduction-

Star 1: all signal grounds- input jacks, volume control, output tube grid resistors

Star 2: all cathode bypasses, output grounds & other non-power supply grounds

Star 3: all power supply grounds & PS capacitor negatives

These 3 stars should then be connected together at one point, and connected to the chassis. Also you could try shielded cable in the input section, with the shield attached to the signal star ground. The filament power supply should be grounded at one end or with resistors as someone else said.

Also have you tried swapping the preamp tubes for some other ones? Even swapping sides- you could see whether the hum folloes the valves- they might be noisy...

I also suggest connecting a battery powered source like an ipod or similar to make sure you don't have a simple earth loop. Also try moving it away from other mains powered gear.

Don't give up! You will surely solve it with a little detective work :)
 
Wow, a lot of great advice. Thanks!

EC and SY: I'll get a hi-rez shot of my amp up tonight. Here is a link to a picture of someone else's if people are need to get an idea of the layout. Mine is laid out in almost identical fashion.

martinab2: The initial problem posted on the AA thread was a real lawnmower hum caused by a loose ground connection - it has since been resolved. The issue presently at hand measures small, but is still too loud to be acceptable IMO.

testlab: Heater voltages on the stock ELLA are 6.3 VAC. If I were to run 100 ohm resistors from the filaments to ground, would that be for all valves or just the output section? driver section?

ray: I have yet to acquire an o-scope, so I can't say for sure what frequency I'm dealing with. There is plenty of noise coming from my tweeter as well as the woofer. Could 60/120hz hum find its way up there?

shifty: The ELLA uses a ground bus down the center of the chassis (see attached pic). When I've been testing it, the amp hasn't been connected near any other gear, so I feel it's safe to say the problem is within the amp itself.

Thanks again fellas, I'll post a picture later tonight.
 
One of the worst offenders is input connectors, such as RCA jacks, that ground to the chassis. I've gotten some wicked nasty hums from that. Now, it may not seem such a big deal having a couple of ohms across a chassis, but it is. Take those few mV of PS hum, add 200X or so open loop gain, and BZZZZZZZ!!!!. Make certain that your input jacks are grounded at one point only: I always return them to the same place on the circuit board where the input stage cathode (emitter, source) ground return is located.
 
Was the area around them scraped clean of paint prior to mounting the input jacks? There also appears to be a very large path between the ground reference for the input tubes and the input jacks. You might consider putting the ground tabs back on the input jacks, making that point your chassis ground reference and lifting the one you currently have.
 
The chassis around the input jacks should still have a coat of paint and primer, but it's possible that bare metal might be contacting one of them within the mounting hole itself. I've checked once, but I'll double check again tomorrow. I'll switch the chassis grounding point as well.
Thanks.
 
Can't stress that too highly- there should only be two places where ANYTHING contacts the chassis, the safety ground (the green wire from the power cord) and the circuit ground, which absolutely should be at the input. The power supply ground should be at the far end of the ground bus, and the ground bus should be directly connected to the PS cap's negative terminal. Since the transformer-rectifier-cap ripple currents are high, even moving the ground bus a short distance from the cap terminal along that loop will cause surprising amounts of hum.

If the hum goes away with nothing plugged into the inputs but reappears when something is plugged in, then the signal ground should be lifted from the chassis. It can be referenced back to chassis ground through a small (10-30 ohm) resistor, which breaks up the ground loop.
 
Sorry tokyofist, I was fooled by the use (very scary IMO) of RCA jacks as bias sample points. On closer look it appears the designer chose a point on the input sub-plate, beneath the paired Solens in the diyhifisupply pic, as the signal ground-to-chassis point. I don't see it in your amp though. Where does your signal ground connect to the chassis?
 
Ex-Moderator
Joined 2003
tokyofist said:
The ground bus connects to the chassis at one of the mounting screws for the power supply PCB. Look in the top view picture.

Well, that'll be your problem, then. When you said you had hum with buzz coming out of the tweeters I thought it had to be power supply ripple getting where it shouldn't. Lift the chassis connection from the power supply PCB and make a new chassis connection from the end of your 0V bus bar that's furthest from the power supply. At the same time, make sure there aren't any other connections to chassis.

Another thing to attend to is the heater wiring. I can't see any connection from the 6.3V to chassis, unless there's a transformer centre tap that I can't see. Connect a 47R resistor from each side of the 6.3V directly to chassis at the first valve where the 6.3V leaves the transformer. If you're feeling posh, add 100nF in parallel with each 47R.

Make sure you use a serrated washer between your earth tag and chassis and that you do the screw up tight. I use M6 (that's 1/4" to you) and do it up with a spanner. You're looking for a low-resistance bond that stays that way for years. Scrape the chassis clean of paint or anodizing before attempting to make a chassis bond. Use this method for the mains safety earth and any other connections to the chassis.
 
You are all gods among men. I first swapped the chassis grounding point to a location near the inputs. This had no effect, however. I then added a resistor and cap in parallel from each 6.3V winding to ground. Bingo :smash: :smash: :smash:! No more noise, even with my ear at the speaker.

I owe you all a beer. We'll of course figure something else out for the teetotalers. I should have come here two months ago and saved myself much agony and frustration.

Thanks much again, although I'm sure I'll be back for future projects.
 
Congrats! :) Another good trick if the filament secondary has a centre tap is to create a resistive divider from B+ to ground - the series string doesn't need to draw much current - bypass the lower resistor to ground with a cap and tie the filament secondary centre tap to the positive end of the cap. Choose a centre point voltage that comfortably exceeds the maximum positive swing any cathode will see. Also make sure not to exceed the heater-filamant rating for any connected tube. This biases the filament positive relative to the cathode and effectively elminates the possibility of the first inducing noise on the second.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.