I see some commercial amp with parallel tube, but I have never seen a DIY, or it's not popular. What's the problem if I put two tubes in parallel ?
Thanks,
Davide
Thanks,
Davide
The main problem is proper load sharing between the devices. Tolerances of the tube parameters, such as transconductance, can be quite high. Without special precautions, one tube will bear more of the load current than the others, and it will fail prematurely.
If you use fixed bias, you need current sharing resistors (lossy), and it is a must to use separate bias controls for each tube, and then balance the tubes very carefully against each other. With cathode bias, the situation is a bit better, but still it would be good to use matched tubes.
For DIY, it's often easier to choose a single, bigger tube.
If you use fixed bias, you need current sharing resistors (lossy), and it is a must to use separate bias controls for each tube, and then balance the tubes very carefully against each other. With cathode bias, the situation is a bit better, but still it would be good to use matched tubes.
For DIY, it's often easier to choose a single, bigger tube.
Depends on usage. For MI I often mix output tubes deliberately to get the raw sound and easy overdrive. Kavermei is quite right: the tube specs can vary so much, nearly 2:1. Choose a vendor who will match. If you are using an E&I output core, it can accept more out of balance primary currents from each halve without ill effects.
With parallel output stages, higher power output for half the B+ voltage compared to single p-p stage: the primary load impedance is halved, requiring an easier designed output transformer with increased bandwidth. ( Tranny designers like this).
Drawbacks; power supply has to emphasise more current, bigger chassis, cheaper electroytics. The plus-minus list goes on.
I am an addict of p-p parallel pairs as I find these amps sound louder.
richy
With parallel output stages, higher power output for half the B+ voltage compared to single p-p stage: the primary load impedance is halved, requiring an easier designed output transformer with increased bandwidth. ( Tranny designers like this).
Drawbacks; power supply has to emphasise more current, bigger chassis, cheaper electroytics. The plus-minus list goes on.
I am an addict of p-p parallel pairs as I find these amps sound louder.
richy
My Edcor CXPP30-MS-5Ks arrived at work today. Nice heavy chunks of iron.
I plan of doing a parallel push pull using 6P1P-EVs with it. The transformers are way over-rated for the application (30W rating for a 16-20W amp), but it's an experiment so I figure a bigger transformer is better as I can use it for something else if I'm not satisfied with the PPP 6P1P.
I'll experiment with Cathode bias, LED bias and maybe Current Source Bias.
I have to finish the 6P1P PP build and get it up stairs for the wife first though, so it will be a while till I do anything with parallel push pull.
I plan of doing a parallel push pull using 6P1P-EVs with it. The transformers are way over-rated for the application (30W rating for a 16-20W amp), but it's an experiment so I figure a bigger transformer is better as I can use it for something else if I'm not satisfied with the PPP 6P1P.
I'll experiment with Cathode bias, LED bias and maybe Current Source Bias.
I have to finish the 6P1P PP build and get it up stairs for the wife first though, so it will be a while till I do anything with parallel push pull.
Gimp: The catch: You mention power rating but at which lowest design frequency ? A -3dB LF cut off at 40Hz is nearly half the iron size of one designed for -3dB at 20Hz cutoff.
The practical handling limitations of 2x 20Hz 150W lumps and a mains transformer requires needs muscle and no trip ups..
There's another thread allied on this: "choosing an output transformer".
Good luck
richy
The practical handling limitations of 2x 20Hz 150W lumps and a mains transformer requires needs muscle and no trip ups..
There's another thread allied on this: "choosing an output transformer".
Good luck
richy
Thanks richwalters.
I've been following the thread, but I'll go back and look at it again.
I have no need of anything over about 30W and am actually quite pleased with the 8W I get from the current 6P1P amp. I just like to experiment.
I've been following the thread, but I'll go back and look at it again.
I have no need of anything over about 30W and am actually quite pleased with the 8W I get from the current 6P1P amp. I just like to experiment.
Can you post the schematic of the 6P1P, these are penthode with 9 pin socket. I think I bought two by mistake, maybe I can put them at work 🙂
Coming back to parallel tubes, I understand that I can take any SE schematic, let's say with cathode bias, connect anodes and grids together, play with the value of cathode resistor to match the current in the tubes, and it will work ? PSU and OPT will see one tube with double power and current. I have to halves the load ? so if I use KT88, that likes 5K, I should use 2.5K ?
D.
Coming back to parallel tubes, I understand that I can take any SE schematic, let's say with cathode bias, connect anodes and grids together, play with the value of cathode resistor to match the current in the tubes, and it will work ? PSU and OPT will see one tube with double power and current. I have to halves the load ? so if I use KT88, that likes 5K, I should use 2.5K ?
D.
The 6P1P (-e/-EV/-B etc) are nine pin Noval tubes(B9A Socket). The latest schematic is here:
http://www.diyaudio.com/forums/tubes-valves/152614-sven-6p1p-pp-amp-5.html
The 6P1P is actually a Beam Power Tetrode (with the beam forming electrode directly connected to the cathode internal to the tube) not a pentode.
Only two screens available, (1) Grid, and (2) Screen.
Yes, when connecting two tubes in parallel you will cut the effective plate resistance and need to reduce the matching impedance by a factor of two.
I am using a 5K transformer with two 6P1Ps in parallel on each side push-pull. They normally use a 10K transformer with only two tubes driving push-pull.
I'm not sure how it would work as a SE, but that sounds about right. You could improve the current matching by using separate cathode resistors that you could adjust to set the current.
http://www.diyaudio.com/forums/tubes-valves/152614-sven-6p1p-pp-amp-5.html
The 6P1P is actually a Beam Power Tetrode (with the beam forming electrode directly connected to the cathode internal to the tube) not a pentode.
Only two screens available, (1) Grid, and (2) Screen.
Yes, when connecting two tubes in parallel you will cut the effective plate resistance and need to reduce the matching impedance by a factor of two.
I am using a 5K transformer with two 6P1Ps in parallel on each side push-pull. They normally use a 10K transformer with only two tubes driving push-pull.
I'm not sure how it would work as a SE, but that sounds about right. You could improve the current matching by using separate cathode resistors that you could adjust to set the current.
Shadetree impedence cut --
Remember that there is a "shadetree" (i.e, uncivilized) solution to this problem of paralleling tubes IF you have multiple impedence taps on your OT *AND* the OT can support the addtional wattage of the second SE tube.
For example, lets say you have a SE 6v6 that wants 5K primary impendence on and OT with 4, 8, and 16 taps into an eight ohm speaker. If you parallel a second 6v6 then you want a 2k5 primary impedence. You can accomplish this in a very crass, uncivilized (and cheeaaap) way by connecting your 8 ohm speaker into the 4-ohm secondary tap.
WARNING: This is mostly a quicky breadboarding trick. Don't try this with your LaScallas or anything like that. Run the amp a while and check the OT temp. If its much hotter than usual - better get a higher wattage OT. But at least you can giver a quick listen. You MUST observe wattage limits - two tubes put out more watts than one (not really twice as much because they must *share* the same input signal level after all -- but more than one).
Kinda feel guilty even bringing this up after following the "Choosing an Output Transformer" thread. It humbled me to say the least. Frequency concerns are real -- it may not sound very good - but if its just for a lil ole junk guitar amp (my usual victim) - give it a rip (hopefully not literally <grin>).
If I got this backwards anybody -- please set me straight.
Remember that there is a "shadetree" (i.e, uncivilized) solution to this problem of paralleling tubes IF you have multiple impedence taps on your OT *AND* the OT can support the addtional wattage of the second SE tube.
For example, lets say you have a SE 6v6 that wants 5K primary impendence on and OT with 4, 8, and 16 taps into an eight ohm speaker. If you parallel a second 6v6 then you want a 2k5 primary impedence. You can accomplish this in a very crass, uncivilized (and cheeaaap) way by connecting your 8 ohm speaker into the 4-ohm secondary tap.
WARNING: This is mostly a quicky breadboarding trick. Don't try this with your LaScallas or anything like that. Run the amp a while and check the OT temp. If its much hotter than usual - better get a higher wattage OT. But at least you can giver a quick listen. You MUST observe wattage limits - two tubes put out more watts than one (not really twice as much because they must *share* the same input signal level after all -- but more than one).
Kinda feel guilty even bringing this up after following the "Choosing an Output Transformer" thread. It humbled me to say the least. Frequency concerns are real -- it may not sound very good - but if its just for a lil ole junk guitar amp (my usual victim) - give it a rip (hopefully not literally <grin>).
If I got this backwards anybody -- please set me straight.
Another thing to remember when parallelling tubes is, that the (combined) grid leak resistance must be made accordingly smaller and therefore it must also be checked that the driver stage can supply enough power to drive it. Whether or not this is a problem, depends on the type of output tube... Some are easier to drive than others.
Moonbird, I talked to someone at Edcor (don't remember her name) before ordering the transformers as that is exactly what I want to be able to do. The secondary windings are all the same heavy guage wire, so there should be no issue with overheating the secondary, or increased resistive loss. Some transformers I've disected used smaller guage wire for the higher Z taps.
If the primary windings determine the primary inductance (critical for LF response), and the ratio is determined by the number of secondary turns, then I would expect to be able to run the Eddy with 10K PP Tubes and use the 16 ohm tap with an 8 ohm speaker. According to the person at Edcor I spoke with this will be fine.
My goal is to build a 6P1P parallel PP, and be able to pull half the tubes, switch speaker connections and make comparisons.
I plan on running a seperate grid resistor and coupling cap for each driver tube.
If the primary windings determine the primary inductance (critical for LF response), and the ratio is determined by the number of secondary turns, then I would expect to be able to run the Eddy with 10K PP Tubes and use the 16 ohm tap with an 8 ohm speaker. According to the person at Edcor I spoke with this will be fine.
My goal is to build a 6P1P parallel PP, and be able to pull half the tubes, switch speaker connections and make comparisons.
I plan on running a seperate grid resistor and coupling cap for each driver tube.
Good to know
The more I hear about Edcor the more I think my days of scouring the earth for old 50s OTs is over .. thx for the info!
The more I hear about Edcor the more I think my days of scouring the earth for old 50s OTs is over .. thx for the info!
Yes, backwards. If you want a 2.5K load out of a 5K transformer put the 8 ohm load on the 16 ohm tap.
It isn't such an uncivilized method after all. I do it all the time. That said there are limitations. The primary must have enough inductance to support the new impedance, which is usually the case when "ratioing down" (using the transformer for less than its rated impedance). The losses in the OPT will usually increase since you are now stuffing twice the DC current through the same winding resistance. Observe the maximum DC current rating in SE OPT's, or at least don't grossly violate it. Magnetic saturation will usually limit the maximum power. It is usually a function of frequency so you can get much more than the rated power through a transformer if you limit the lower frequencies. This trick usually works better with P-P OPT's than SE (due to saturation), and is usually useful for only one "step" (divide or multiply impedance by 2). Examples:
I am using a pair of transcendar 3K ohm SE OPT's designed for 300B's at 6K in my Simple SE amp with KT88's. They work very well and sound excellent.
I have tried One Electron 3K ohm SE OPT's at 1500 ohms, They are quite lossy due to their high DC resistance, and pretty useless at 750 ohms.
I have a bunch of super cheap P-P 6600 ohm OPT's designed for guitar amp use. These work far better than they should based on their budget construction. They work good at 3300 ohms and OK at 13.2K ohms. These were specified for "80 watts" in guitar amp duty (80 Hz minimum). For Hifi use down to 25 Hz they handle about 30 watts. Using them at 3300 ohms reduces this to about 25 watts. Measured results in a 6CW5 based Simple P-P (3300 ohm) shows 25 watts at 25 Hz with 3% distortion. Magnetic saturation reduces this to about 18 watts at 20 Hz. I have however cranked over 150 watts through these same OPT's at 3300 ohms at 1KHz!
I have a pair of monster Plitron P-P OPT's rated for 400 watts. They are designed for 1250 ohm use with 2, 4, and 8 ohm taps. The original application was for a Marshall bass guitar amp. They work just great for HiFi at 2500 and 5000 ohms. They will do 10 Hz to 50 KHz at any power level that I have been able to feed them up to 200 watts.
Hmm -- well I certainly misunderstood the old shadetree who originally told me this one -- but I certainly believe you Tubelab -- you have done it all. ANNND it could go a long way toward explaining the acrid smoke that blows through my "lab" from time to time 🙂.
At the risk of being inappropriate, this is one situation that I had hoped might be addressed when I started the "Choosing an Output Transformer" thread. Is there anyway to detect a good "budget" OT standing at a hamfest table far from your bench (I have had to face the fact I can't buy them all <grin>)?? Tubelab, could I invite you to comment on this in that thread ... I know it might be a little awkward given how that discussion has taken such an "orthodox" path. My sincere apologies to this threads author.
Last edited:
No need to apologize, actually I found here answer to some questions I had in mind. I was also following the discussion on the other thread, but it does not address the question "How to choose an output transformer, (I would add) when you are in the shop". I find very difficult to find real specs and technical details.
I hate that I end up choosing the cutest.
As I am never sure of whet I want or need I try to buy transformers that have multiple taps, although I see that the very high end one and single primary.
So it means that I and use the Hashimoto 5/7k (max 100mA) at 3.5k shifting place on the secondary. I am thinking about my Tubelab SE with 300B.
Once the salesman told me "it affects the frequency response", but I could not understand why.
Thanks,
Davide
I hate that I end up choosing the cutest.
As I am never sure of whet I want or need I try to buy transformers that have multiple taps, although I see that the very high end one and single primary.
So it means that I and use the Hashimoto 5/7k (max 100mA) at 3.5k shifting place on the secondary. I am thinking about my Tubelab SE with 300B.
Once the salesman told me "it affects the frequency response", but I could not understand why.
Thanks,
Davide
- Status
- Not open for further replies.