Hi guys,
I'm relatively new to the world of tubes having only previously diy'ed a pre-amp (Simple 5687 DIY Paradise amp) and a triode strapped SE EL34 power amp.
This weekend I completed the Morgan Jones Bevois valley Ultralinear EL84 Push Pull amplifier which kicked off my current problem which I would appreciate some help with.
Firstly some background. I built the EL84 PP as per Morgan Jones schematic using Hammond iron 370HX + 1608 output tx's (similar to Kofi's build) except I changed the gridstoppers on the E88CC from 330R to 4k7 (high but all I had to hand) on both pins 2 & 7. This appeared to overcome all the stabilisation issues.
I tested this amp using a MP3 player connected directly and all was well except I had to use nearly full volume on the MP3 player due to the low sensitivity of the EL84 Push Pull amp. OK I was happy at this stage as I can 'simply' alter the feedback resistor/cap set-up to increase gain as suggested on an earlier thread.
Anyway, as I already had the 5687 pre-amp available I decided to use this between the MP3 player and the EL84 PP amp. This worked fine with volume control at ~12 o'clock but this was not unexpected.
However, when checking the amount of background noise (no music playing) I noticed that the loudspeaker cone was intermittently flapping very excessively so much that I quickly turned the amps off fearing damage to the drive units
On further investigation, I discovered that this 'flapping' only occurred when the pre-amp was connected in the chain. With just MP3 player & EL84 Push Pull amp connected together (no music playing) or EL84 Push Pull amp switched on with no source connected the loudpeaker cone didn't move at all.
By the way my loudspeakers are DIY using the Fostex full range driver FE167E in their recommended bass reflex cabinets without using any crossover.
I guess the excessive movement is due to some DC voltage on the output to the loudspeaker but I don't understand why it is happening only when the pre-amp is connected? BTW I need to use the pre-amp when I'm playing vinyl so I will need a solution.
I gave up on trying to fix this a few hours ago as I'm pretty clueless about the technicalities of tube amps so hopefully someone here can suggest a really simple solution which a beginner like me can implement.
Any help is appreciated
Colin.
I'm relatively new to the world of tubes having only previously diy'ed a pre-amp (Simple 5687 DIY Paradise amp) and a triode strapped SE EL34 power amp.
This weekend I completed the Morgan Jones Bevois valley Ultralinear EL84 Push Pull amplifier which kicked off my current problem which I would appreciate some help with.
Firstly some background. I built the EL84 PP as per Morgan Jones schematic using Hammond iron 370HX + 1608 output tx's (similar to Kofi's build) except I changed the gridstoppers on the E88CC from 330R to 4k7 (high but all I had to hand) on both pins 2 & 7. This appeared to overcome all the stabilisation issues.
I tested this amp using a MP3 player connected directly and all was well except I had to use nearly full volume on the MP3 player due to the low sensitivity of the EL84 Push Pull amp. OK I was happy at this stage as I can 'simply' alter the feedback resistor/cap set-up to increase gain as suggested on an earlier thread.
Anyway, as I already had the 5687 pre-amp available I decided to use this between the MP3 player and the EL84 PP amp. This worked fine with volume control at ~12 o'clock but this was not unexpected.
However, when checking the amount of background noise (no music playing) I noticed that the loudspeaker cone was intermittently flapping very excessively so much that I quickly turned the amps off fearing damage to the drive units
On further investigation, I discovered that this 'flapping' only occurred when the pre-amp was connected in the chain. With just MP3 player & EL84 Push Pull amp connected together (no music playing) or EL84 Push Pull amp switched on with no source connected the loudpeaker cone didn't move at all.
By the way my loudspeakers are DIY using the Fostex full range driver FE167E in their recommended bass reflex cabinets without using any crossover.
I guess the excessive movement is due to some DC voltage on the output to the loudspeaker but I don't understand why it is happening only when the pre-amp is connected? BTW I need to use the pre-amp when I'm playing vinyl so I will need a solution.
I gave up on trying to fix this a few hours ago as I'm pretty clueless about the technicalities of tube amps so hopefully someone here can suggest a really simple solution which a beginner like me can implement.
Any help is appreciated
Colin.
Hi Colin
I understand your feeling quite well (I also lost weekends troubleshooting things), so, I will post something for you to reflect upon, but surely other people will jump in!
Looks like motorboating (valve amplifiers p410-411). It is just weird, as this problem is mostly caused by feedback through the power supply, but I suppose you are using separate PS for the pre with 5687 and the Bevois Valley power.
If you can not solve the problem, I recommend you to describe the PS arrangements you made!
Good luck, Erik
I understand your feeling quite well (I also lost weekends troubleshooting things), so, I will post something for you to reflect upon, but surely other people will jump in!
Looks like motorboating (valve amplifiers p410-411). It is just weird, as this problem is mostly caused by feedback through the power supply, but I suppose you are using separate PS for the pre with 5687 and the Bevois Valley power.
If you can not solve the problem, I recommend you to describe the PS arrangements you made!
Good luck, Erik
Hi,
Thanks for the replies.
To answer the questions raised.
Burnedfingers - The capacitor on cathode resistor of 5687 is 470mfd bypassing a 470R. (See attached schematic)
Sy - Your description of a 'grunting pre-amp' sounds about right! I'm afraid I don't have an scope available only a DVM so I guess solving the problem maybe somewhat harder than it should be. Schematic for the 5687 pre-amp is attached (taken from DIY Paradise website.)
My power supply for this pre-amp is the same as the schematic except I used a 350-0-350 power trafo (Hammond 373X) & 5Y3 rectifier valve. Also (but probably not relevant here) I switched to DC heater supply for the 5687 due to hum. Only other changes 0.1uF caps were put on the input RCA's.
One final point I don't recall the loudspeaker cone flapping when I connect the 5687 pre-amp with my EL34 SE power amp.
Cheers
Colin.
Thanks for the replies.
To answer the questions raised.
Burnedfingers - The capacitor on cathode resistor of 5687 is 470mfd bypassing a 470R. (See attached schematic)
Sy - Your description of a 'grunting pre-amp' sounds about right! I'm afraid I don't have an scope available only a DVM so I guess solving the problem maybe somewhat harder than it should be. Schematic for the 5687 pre-amp is attached (taken from DIY Paradise website.)
My power supply for this pre-amp is the same as the schematic except I used a 350-0-350 power trafo (Hammond 373X) & 5Y3 rectifier valve. Also (but probably not relevant here) I switched to DC heater supply for the 5687 due to hum. Only other changes 0.1uF caps were put on the input RCA's.
One final point I don't recall the loudspeaker cone flapping when I connect the 5687 pre-amp with my EL34 SE power amp.
Cheers
Colin.
Attachments
SY said:
I was wondering ...... would it be appropriate, since no 'scope is available, to connect the preamp to a soundcard running TrueRTA or similar? Or is it too risky? How about using test signals and a DMM to check for high voltages first?
I'd be really interested to know IF it's a good idea or not.
Regards,
Martin
If the problem is very low in frequency, it will be tough to see on a sound-card. From the description, it seems like it's 2-4Hz and probably shaky. A scope is better for this sort of thing.
Leaky cathode bypass might be the problem- you can diagnose it by removing the cap altogether (this will drop gain) and seeing if the problem remains. You could also be suffering from line voltage variations, which will not be well rejected by a passive filter, and will also make the heater voltage bounce around a bit.
The general question of sond-card input protection is a good one; I tend to err toward the scared-to-death side. Between the DUT and the sound-card, I'll use a buffer which will limit the voltages seen by the sound-card. You need something like that anyway or the low input Z of the sound-card (typically 10k) will load down the circuit you're trying to measure.
Leaky cathode bypass might be the problem- you can diagnose it by removing the cap altogether (this will drop gain) and seeing if the problem remains. You could also be suffering from line voltage variations, which will not be well rejected by a passive filter, and will also make the heater voltage bounce around a bit.
The general question of sond-card input protection is a good one; I tend to err toward the scared-to-death side. Between the DUT and the sound-card, I'll use a buffer which will limit the voltages seen by the sound-card. You need something like that anyway or the low input Z of the sound-card (typically 10k) will load down the circuit you're trying to measure.
colinB said:
I tested this amp using a MP3 player connected directly and all was well except I had to use nearly full volume on the MP3 player due to the low sensitivity of the EL84 Push Pull amp. OK I was happy at this stage as I can 'simply' alter the feedback resistor/cap set-up to increase gain as suggested on an earlier thread.
However, when checking the amount of background noise (no music playing) I noticed that the loudspeaker cone was intermittently flapping very excessively so much that I quickly turned the amps off fearing damage to the drive units
On further investigation, I discovered that this 'flapping' only occurred when the pre-amp was connected in the chain. With just MP3 player & EL84 Push Pull amp connected together (no music playing) or EL84 Push Pull amp switched on with no source connected the loudpeaker cone didn't move at all.
I guess the excessive movement is due to some DC voltage on the output to the loudspeaker but I don't understand why it is happening only when the pre-amp is connected? BTW I need to use the pre-amp when I'm playing vinyl so I will need a solution.
Colin.
Any chance of a circuit diagram ?
richj
Hi,
Just an update regarding the ongoing problem I'm having with both loudspeaker cones 'flexing' when I link my 5687 pre-amp + EL84 PP power amp together (even when there is nothing connected to the input). Spent a couple of hours playing this evening following some of the earlier suggestions.
A quick review; the problem only appears to occur when this combination of amps is used. For example,
1. EL84 PP power amp only connected to speakers - both cones are static.
2. I combined the EL84 PP power amp together with a DIY solid state pre-amp (ESP P37) and nothing on the pre-amp input and both speaker cones were static.
3. Combined 5687 pre-amp + triode connected EL34 SE power amp and nothing connected to pre-amp input and again both speaker cones were static.
4. Replaced valves on the pre-amp with spare set (5Y3 + 5687) & reconnected to EL84 PP power amp + nothing on pre-amp input - still seeing cone flexing on both speakers.
5. Disconnected one of the cathode bypass capacitors on the 5687 tube - left the other 470mfd in place. Coupled pre-amp back up to EL84 PP amp with no signal on pre-amp input but still seeing cone movement on both speakers.
6. Increased the bypass capacitor value from 470mfd to 1000mfd on one side of the 5687only (same side as for #5). Coupled pre-amp back up to EL84 PP amp with no signal on pre-amp input but still seeing cone flexing on both speakers.
So now I'm back to square one - anyone got any further suggestions?
What I don't get is why the 5687 preamp + EL34 SE combo doesn't give problems but the 5687 preamp + EL84 PP amp does but then again I'm only a chemist!
Rich: I couldn't attach a schematic of the EL84PP amp for fear of copyright issues. If you have Morgan Jones 'Valve Amplifiers' 3rd edition the schematic is on p 454. I didn't use a 285V HT regulator but a choke(20H)/capacitor(100uF) combination suggest on post #109 of a thread "Kofi Annan in: "Push and Pull with Me" (I'm not clever enough to know how I link to this!)
Otherwise I built the EL84 power amp as per schematic but needed more than 330R as grid stoppers on E88CCs - I finished at 4k7R but this is fair from optimised.
Well any further help would be appreciated.
Thanks
Colin.
Just an update regarding the ongoing problem I'm having with both loudspeaker cones 'flexing' when I link my 5687 pre-amp + EL84 PP power amp together (even when there is nothing connected to the input). Spent a couple of hours playing this evening following some of the earlier suggestions.
A quick review; the problem only appears to occur when this combination of amps is used. For example,
1. EL84 PP power amp only connected to speakers - both cones are static.
2. I combined the EL84 PP power amp together with a DIY solid state pre-amp (ESP P37) and nothing on the pre-amp input and both speaker cones were static.
3. Combined 5687 pre-amp + triode connected EL34 SE power amp and nothing connected to pre-amp input and again both speaker cones were static.
4. Replaced valves on the pre-amp with spare set (5Y3 + 5687) & reconnected to EL84 PP power amp + nothing on pre-amp input - still seeing cone flexing on both speakers.
5. Disconnected one of the cathode bypass capacitors on the 5687 tube - left the other 470mfd in place. Coupled pre-amp back up to EL84 PP amp with no signal on pre-amp input but still seeing cone movement on both speakers.
6. Increased the bypass capacitor value from 470mfd to 1000mfd on one side of the 5687only (same side as for #5). Coupled pre-amp back up to EL84 PP amp with no signal on pre-amp input but still seeing cone flexing on both speakers.
So now I'm back to square one - anyone got any further suggestions?
What I don't get is why the 5687 preamp + EL34 SE combo doesn't give problems but the 5687 preamp + EL84 PP amp does but then again I'm only a chemist!
Rich: I couldn't attach a schematic of the EL84PP amp for fear of copyright issues. If you have Morgan Jones 'Valve Amplifiers' 3rd edition the schematic is on p 454. I didn't use a 285V HT regulator but a choke(20H)/capacitor(100uF) combination suggest on post #109 of a thread "Kofi Annan in: "Push and Pull with Me" (I'm not clever enough to know how I link to this!)
Otherwise I built the EL84 power amp as per schematic but needed more than 330R as grid stoppers on E88CCs - I finished at 4k7R but this is fair from optimised.
Well any further help would be appreciated.
Thanks
Colin.
I'm a chemist, too, but despite that, I've managed not to electrocute myself yet.
It is entirely possible that the frequency response of the EL34 does not go down to the subsonic. You're putting garbage in, but it's not making it to the output. it still shouldn't be there. I'd focus on maybe changing the 5687 (could be heater-to-cathode leakage if the replacement came from the same batch) and making sure that the supplies (including heater) are nice and stable. How are you powering the heater anyway? How is it referenced to DC ground?
edit: Is the cone motion correlated between the two channels?
It is entirely possible that the frequency response of the EL34 does not go down to the subsonic. You're putting garbage in, but it's not making it to the output. it still shouldn't be there. I'd focus on maybe changing the 5687 (could be heater-to-cathode leakage if the replacement came from the same batch) and making sure that the supplies (including heater) are nice and stable. How are you powering the heater anyway? How is it referenced to DC ground?
edit: Is the cone motion correlated between the two channels?
SY said:
To me etwas a bit unclear. However I did come across an amp with a sim problem and the diagnosis was somewhat different. Worth a mention.
The orig start of this thread describes quote; << as I can 'simply' alter the feedback resistor/cap set-up to increase gain as suggested on an earlier thread.>> (this is risky practise in a hifi power amp and quite alot of circuits are about).
That to me seems like the negative feedback loop in the power amp can/has be conveniently used to alter gain and the more feedback the greater the risk of oscillation. This changes the closed loop response of the amp. The LF motorboating effect can be caused by too high value coupling caps typ to driver stages. This creates the nfb loop to hold up the LF response way down to subsonic LF eventually bordering positive feedback i.e oscillation and a slight aggravation can trigger this.
The symtoms may appear sim to psu problems, which can be confusing and also cause this.. but in low power amps which spend most of their time in class A, the LF motorboating tendency will cause the o/p stage operating point to shift into class B causing a drop in the B+ creating more problems as the volts in the earlier stages falls.
This analysis may seem awkard for tube newcomers but does happen in practise and is easy to blame the psu.
The AB test is important. See if the other channel does it. It may not simply oscillate as slight variations in component values may "just be off" oscillation but if this channel can also be made to show the same symptoms, then something more serious is up.
Sonically, amps which have very low LF response also have poor hangover tendencies on transient recovery, the sound quality appears muddled.
There is a "battery test" for this which I shall describe later.
richj
Hi,
Just a few responses to questions posed by SY
I'm only using the cheap Hammond single ended output trafo's (125GSE) on the EL34 amp that have limited low end frequency capability so I suspect you are right with this.
So as you say I should concentrate on the pre-amp.
I swapped the 5687 valve last night with a spare and there was still cone movement. These valves came from two different vendors so I think it is very unlikely that they came from the same batch. I can check the boxes when I get home (I'm in work now!)
OK, this is roughly what I did.
I used the 6.3VAC from the power trafo (373X) into 31DQ06 schottky diodes bridge then filtered as 10000microfarad cap/1r/10000microfarad cap/1r/10000microfarad cap which gave 8.9VDC without the valve in place.
With 5687 installed this reduced to 7VDC. This is too high but with 0.5R in both 'feeds' to the valve heater the voltage dropped to about 6.2VDC.
The centre tap is connected to ground at the power supply cap.
Info is in this thread.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=82846&perpage=15&pagenumber=2
Sy: If you think it would be a good diagnostic test I could go back to AC on the 5687 heater supply? Maybe I'll post a pic of the underside of the pre-amp later which may show something obviously wrong but I should warn you it ain't a pretty sight
I'm gonna need alot of help with this - I assume you mean HT regulation here. That's a whole new concept for me
I tried to assess this last night but couldn't be sure. First impressions were that the movement was correlated but after watching for longer I was less sure. Hey I'll check this again to-night and let you know.
Rich: Thanks for your reply.
I should stress that I haven't adjusted any feedback settings on the EL84 PP Power amp. They remain as per Morgan Jones schematic. I think I was a bit hasty with this statement. Sorry if it confused.
Thanks for the help
Cheers
Colin.
Just a few responses to questions posed by SY
I'm only using the cheap Hammond single ended output trafo's (125GSE) on the EL34 amp that have limited low end frequency capability so I suspect you are right with this.
So as you say I should concentrate on the pre-amp.
I swapped the 5687 valve last night with a spare and there was still cone movement. These valves came from two different vendors so I think it is very unlikely that they came from the same batch. I can check the boxes when I get home (I'm in work now!)
OK, this is roughly what I did.
I used the 6.3VAC from the power trafo (373X) into 31DQ06 schottky diodes bridge then filtered as 10000microfarad cap/1r/10000microfarad cap/1r/10000microfarad cap which gave 8.9VDC without the valve in place.
With 5687 installed this reduced to 7VDC. This is too high but with 0.5R in both 'feeds' to the valve heater the voltage dropped to about 6.2VDC.
The centre tap is connected to ground at the power supply cap.
Info is in this thread.
http://www.diyaudio.com/forums/showthread.php?s=&threadid=82846&perpage=15&pagenumber=2
Sy: If you think it would be a good diagnostic test I could go back to AC on the 5687 heater supply? Maybe I'll post a pic of the underside of the pre-amp later which may show something obviously wrong but I should warn you it ain't a pretty sight
I'm gonna need alot of help with this - I assume you mean HT regulation here. That's a whole new concept for me
I tried to assess this last night but couldn't be sure. First impressions were that the movement was correlated but after watching for longer I was less sure. Hey I'll check this again to-night and let you know.
Rich: Thanks for your reply.
I should stress that I haven't adjusted any feedback settings on the EL84 PP Power amp. They remain as per Morgan Jones schematic. I think I was a bit hasty with this statement. Sorry if it confused.
Thanks for the help
Cheers
Colin.
If you've got a scope, what you want to do is look at the power supply rails and see if they're rumbling. Use the AC coupled option- most scopes will have a voltage limit there, and if the rail is greater than that, you'll have to do a bit of improvising. Basically, attach a 3 or 4uF polypropylene cap to the rail and the other end to the scope probe (usual warnings: this is a deadly high voltage area, use common sense, basic safety rules, and be aware that the cap will hold several hundred volts of charge). This method substitutes for the series cap built into the oscilloscope.
What you should see is just a bit of residual ripple. If the rail is bouncing up and down, it's time to regulate. The basic two transistor regulator scheme from Valve Amplifiers would work very well here, as would a Maida.
One more thing to look at is heater biasing. Normally, that manifests as hum or noise injection, but... Disconnect the heater center tap from ground, then connect it to a source that's (roughly) 20-30V above ground. This can be done by making a voltage divider out of a series 100k and 10k resistor, connecting it from B+ rail to ground (100k on the "high" end), bypassing the 10k with a moderately big cap (10-100uF/63V), then connecting the formerly-grounded heater center-tap to the junction of the 10k, 100k, and electrolytic. A quick and dirty way to just try this out to see the effect is to take a couple of 9V batteries in series, attach the negative to ground, and attach the heater CT to the positive terminal.
What you should see is just a bit of residual ripple. If the rail is bouncing up and down, it's time to regulate. The basic two transistor regulator scheme from Valve Amplifiers would work very well here, as would a Maida.
One more thing to look at is heater biasing. Normally, that manifests as hum or noise injection, but... Disconnect the heater center tap from ground, then connect it to a source that's (roughly) 20-30V above ground. This can be done by making a voltage divider out of a series 100k and 10k resistor, connecting it from B+ rail to ground (100k on the "high" end), bypassing the 10k with a moderately big cap (10-100uF/63V), then connecting the formerly-grounded heater center-tap to the junction of the 10k, 100k, and electrolytic. A quick and dirty way to just try this out to see the effect is to take a couple of 9V batteries in series, attach the negative to ground, and attach the heater CT to the positive terminal.
Hi SY
Just an update following your suggestions.
Sorry to say I don't have a scope (and unlikely to be able to afford to buy one in the near future.) Sadly I don't know anybody who has one either. So I can't carry out any tests on the residual ripple as you suggested.
I gave this battery test a quick try (leaving my heaters on the DC supply) but unfortunately both speaker cones moved once the 5687 pre-amp & EL84 power amp were connected together. So it doesn't look like the heater bias is the culprit.
I sat for a few minutes watching the speaker cones on both channels and I'm fairly certain that the cone movements correlate. The cone motion appears fairly random and sometimes the cones are static for 30s or more before moving several times in succession!
Anyway, as I don't have a scope, it looks like my only real way to progress is to build a HT regulator (Maida or MJ's p361 3rd edition), connect it temporarily and observe the loudspeaker cones to see if problem is solved.
But I need to do some reading about these regulators first so I can work out a parts list etc.
Anyway thanks for your all your help.
Cheers
Colin.
Just an update following your suggestions.
Sorry to say I don't have a scope (and unlikely to be able to afford to buy one in the near future.) Sadly I don't know anybody who has one either. So I can't carry out any tests on the residual ripple as you suggested.
I gave this battery test a quick try (leaving my heaters on the DC supply) but unfortunately both speaker cones moved once the 5687 pre-amp & EL84 power amp were connected together. So it doesn't look like the heater bias is the culprit.
I sat for a few minutes watching the speaker cones on both channels and I'm fairly certain that the cone movements correlate. The cone motion appears fairly random and sometimes the cones are static for 30s or more before moving several times in succession!
Anyway, as I don't have a scope, it looks like my only real way to progress is to build a HT regulator (Maida or MJ's p361 3rd edition), connect it temporarily and observe the loudspeaker cones to see if problem is solved.
But I need to do some reading about these regulators first so I can work out a parts list etc.
Anyway thanks for your all your help.
Cheers
Colin.
Hi all,
I have the same problem with my 6SN7 cathode follower preamp
trying to troubleshoot for at least two months with no progress.
My conclusion is that it's only up to B+ supply (especially filtering).I used a separate regulated PS for B+ and had no cone shaking.
I also noticed that using the same PS with different topologies
(SRPP ,MU follower etc.) the problem was there.
My PS is : 350 0 350 to 5R4GY to RCLCR and then to two branches of CRC one for each channel.
I've changed cap and R values many times and what I found making significant difference without elliminating the problem is to increase all R values and especially the first R after 5R4.
I'll be watching this thread with interest.
Regards Michael
I have the same problem with my 6SN7 cathode follower preamp
trying to troubleshoot for at least two months with no progress.
My conclusion is that it's only up to B+ supply (especially filtering).I used a separate regulated PS for B+ and had no cone shaking.
I also noticed that using the same PS with different topologies
(SRPP ,MU follower etc.) the problem was there.
My PS is : 350 0 350 to 5R4GY to RCLCR and then to two branches of CRC one for each channel.
I've changed cap and R values many times and what I found making significant difference without elliminating the problem is to increase all R values and especially the first R after 5R4.
I'll be watching this thread with interest.
Regards Michael
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.