|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
97.94.49.170
I recently built a EL84 based push-pull stereo amp, using a Hammond 710 power transformer (375-0-375 VAC, 345 ma). It "sings", emitting a high frequency tone that is NOT present in the DC output. The rectifiers (two, separate rectifier and filter for each side, common power transformer) are 6CA4s. Configuration is LCLC. Placing a steel sheet near to the transformer seems to reduce the amplitude, so I am guessing that it is a magnetic issue.
Any suggestions?
Follow Ups:
You can try adding some mass --- maybe Blu-Tack --- to the lammies and see if that reduces the "singing."
Why is the heater grounded? It seems that using two heater transformers would have been a better design choice.
Not sure if it is related to the singing, but there are likely to be circulating currents through the heater-cathode of the 6CA4's either now or with some ageing.
The same filament transformer is used for the amp (4*EL84, 4*6SN7), because there isn't room for two filament transformers. Grounding the center tap of the filament transformer to the audio tubes is a common practice.
The 6CA4 heater-cathode voltage rating is being exceeded in your design. Is that what you intended?
The heater-cathode voltage limit of the 6CA4 is 500 volts. T1 secondary is 350-0-350 VAC, so the cathode voltage is around 300 VDC with the LCLC filter. Why do you say that the heater-cathode voltage has been exceeded?
Dunacn amps "PSU Designer II" shows no problems.
The cathode voltage will follow the conducting anode voltage, minus the Vak conduction drop. The Hammond 710 has a 375-0-375V rating, with an anode peak of circa 530V, which could go higher depending on your mains AC and whether Hammond rate the 375-0-375 at no load or rated load.One end of the heater will be -4V at the time of peak anode voltage.
It looks to me like each 6CA4 will have its heater-cathode rating repetitively reaching the design-max rating, and probably exceed it. The design-max datasheet rating advise is 'should not be exceeded under the worst probable conditions', and this applies throughout the service life of the valve.
In general, heater-cathode resistance falls with age, and can sometimes be a service life limiting issue.
Edits: 10/14/19
Thanks for the analysis. I was curious, so attached a 11:1 external resistor divider from cathode to ground. Then measured the voltage with an oscilloscope. It confirms your analysis. The peak positive voltage at the cathode measures approximately 610 volts with the amp connected (an approximate 100 ma load).
I ordered a 6.3 VAC 4 amp filament transformer (Hammond 167N6). Will also have to replace the 12 amp filament transformer with a 6 amp version in order to squeeze two transformers into the space the 12 amp version occupies.
Would it be better to let the center tap of the transformer connected to the 6CA4 filaments "float" (leave disconnected) or tie it to 300 VDC?
610V is a little more than I would have thought, but yes in to the risky side of a 500V limit.I'd be inclined to tie the 6CA4 heater to its cathode, but you have two 6CA4 for separate supplies and want to use a common heater, so that doesn't seem appropriate.
Tying the heater to one of the 300Vdc nominal elevated levels would be reasonable and easy, although I would suggest using a buffer resistor to limit any current flow if a heater-cathode short occurred. The buffer resistor should normally only have parasitic currents through it if it is not too large, so perhaps 100k 1W to handle a fault.
I'm a fan of using two series 1N4007 in series with each valve diode plate, as a form of safety insurance if and when a 6CA4 goes bad/gassy. And I'm also a fan of an appropriately selected CT fuse for HT windings, especially if there are multiple transformers fed from a single mains fuse. I've come across enough examples of collateral damage, and have the means to add those protective parts, but that may not be to everyone's taste.
Edits: 10/14/19
when the transformer is "singing?" Or is the "singing" coming solely through the audio?
No singing in the audio from the amp. Amp is clean. It's coming from the case of the power transformer. Holding the transformer's hosing stops the noise.
The amp electronics are in a separate enclosure than the power supply, making it easy to identify the power transformer housing as the source of the noise.
Then you have identified the problem!!!
Hammond 200 series power transformers are pretty famous for this. I use the 300 series when I do Hammonds. Mount the transformer differently, use rubber isolation washers, possibly a vice grip clamp on it to dampen. It is vibration.
Rubber washers will be problematic to implement, because the unit is fully built and the transformer mounting bolts are underneath the 50 uFd capacitors. Worth a try though - the block of wood is tacky looking.
Rubber washers are a good idea. Some amplifiers have had this from the factory. It's not just Hammond transformers that are susceptible.
--------------------------
Buy Chinese. Bury freedom.
I've heard transformers buzz, but I don't think I'm familiar with the effect you're describing. Does it sing even with only a light load, like when the EL84s are removed? You might need to unbolt it from the chassis to be sure this isn't a panel resonance.
--------------------------
Buy Chinese. Bury freedom.
No, with no load (other than the bleeder resistors), the noise is essentially inaudible.
I placed a tight fitting block of wood between the power and filament transformers, and that reduced the noise to an acceptable level. Sort of a half baked "fix", but it will do for now. It's an aluminum chassis. Iv'e always used steel a chassis before, and wonder if that has something to do with the problem.
This is a bit of a classic problem. Just pay mind to the nylon shoulder washers- if they are damaged the bolts will become a magnetic short and will heat up the transformer!
schematic?
what chokes are you using?
Schematic below. Very "classic" design. The Hammond components (power transformer, chokes, filament) are new; L1 and L3 are vintage Military Grade potted units.
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: