How to make an 600 ohm LCR have Zin=1200ohm
Great! So a choke is no problem. I am also thinking of a balanced CF with the same tube. What about a PP 100H (25+25H) or more, centertapped 165 or 330ohm/side? As the DC-current is balanced out, this shouldn´t be a problem, but what about size?
Have not decided about input driver yet, but was thinking of something like below around two quadfilars 1+1+1+1(remember you made a pair of small for my Sony CD) if they are doable. Unfortunately I don´t have a clue about was is possible with the the bi/quadfilars.
I already have two sets of LCR chokes. The frequency sim will not work properly due to the input transformer.
Have not decided about input driver yet, but was thinking of something like below around two quadfilars 1+1+1+1(remember you made a pair of small for my Sony CD) if they are doable. Unfortunately I don´t have a clue about was is possible with the the bi/quadfilars.
I already have two sets of LCR chokes. The frequency sim will not work properly due to the input transformer.
- Attachments
-
- LCR6H30CFbal.asc
- (12.91 KiB) Downloaded 642 times
Brgds
Lars
Lars
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Quadfilar is not the way to go. You would need something very similar to what I did for John swenson.
http://www.intactaudio.com/forum/viewtopic.php?t=566
Have you seen Steve Bench's Drop down models trick for LT spice?
http://www.intactaudio.com/forum/viewtopic.php?t=523
it really makes things easy.
dave
http://www.intactaudio.com/forum/viewtopic.php?t=566
Have you seen Steve Bench's Drop down models trick for LT spice?
http://www.intactaudio.com/forum/viewtopic.php?t=523
it really makes things easy.
dave
Dave,
Didn´t bother check the other solution, instead I did some private brainstorming. No coupling caps but lots of iron. You have already made two of the bifilars for me. So this is what I ended up with today:
Didn´t bother check the other solution, instead I did some private brainstorming. No coupling caps but lots of iron. You have already made two of the bifilars for me. So this is what I ended up with today:
- Attachments
-
- LCR6H30CFbalfreq.asc
- (13.95 KiB) Downloaded 558 times
Brgds
Lars
Lars
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
The IT's loading the D3A and the CT choke in the cathode of the 6H30 are pretty straightforward. It is important to remember that the design I would choose for both of these would be different from the quadfilar design for the SACD since the source Z of the D3A is too high for the "reversed bifilar" behavior they will see.
the 4:1 loading the 6C45's is another beast altogether. I'd prefer not to open that can of worms right now since my plate is full and i'd prefer to stick to things i believe in. (plus many folks out there will gladly fill this need)
dave
the 4:1 loading the 6C45's is another beast altogether. I'd prefer not to open that can of worms right now since my plate is full and i'd prefer to stick to things i believe in. (plus many folks out there will gladly fill this need)
dave
Hey Dave,dave slagle wrote: It is important to remember that the design I would choose for both of these would be different from the quadfilar design for the SACD since the source Z of the D3A is too high for the "reversed bifilar" behavior they will see.
the 4:1 loading the 6C45's is another beast altogether. I'd prefer not to open that can of worms right now since my plate is full and i'd prefer to stick to things i believe in. (plus many folks out there will gladly fill this need)
dave
We left the quadfilar after your previous post. Have never heard of "reversed bifilar behaviour" so please fill me in on that subject.
Forgot to say the D3a is actually E86C/6CM4 but the D3a model comes quite close and where available.
The bifilars are the ones you made for me, to be used between first and second stage of an earlier phono amp:
But will they work properly in the low end in this application?
About the PP-choke and the OPTs, these are std items that you do not need to spill your valuable time on.
Or do you, as I hope, have another solution to get balanced with Zout well below 1k?
Brgds
Lars
Lars
Hi Dave,
As I am currently not using the bifilars from my last post, would it be possible to use the cores from them and make two PP-CHOKE1 bobins in the region of 2*150H, DCR not critical? They will be used balanced with I=10-15mA /side. If doable please quote price.
As I am currently not using the bifilars from my last post, would it be possible to use the cores from them and make two PP-CHOKE1 bobins in the region of 2*150H, DCR not critical? They will be used balanced with I=10-15mA /side. If doable please quote price.
- Attachments
-
- LCR6H30CFbalfreq.asc
- (14.5 KiB) Downloaded 558 times
Brgds
Lars
Lars
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Hey Lars,
"reversed bifilar" means that if you reverse the phase (ground reference)of one of the windings, the capacitance rolls off your top end. In the case of the SACD OT's the design of the transformer was such that the with the proper connection bandwidth was out to the Mhz and with the reversed connection the corner was in the 100K range. These were small transformers designed for low impedance, If you reversed the bifilars you currently have I would expect the top to roll off well into the audio band.
I do not see how the current bifilars that you have will fit into this application, As you mentioned you could re-use the cores and get a new pair of bobbins. And a choke to your needed specs is possible. I'll contact you about pricing.
dave
"reversed bifilar" means that if you reverse the phase (ground reference)of one of the windings, the capacitance rolls off your top end. In the case of the SACD OT's the design of the transformer was such that the with the proper connection bandwidth was out to the Mhz and with the reversed connection the corner was in the 100K range. These were small transformers designed for low impedance, If you reversed the bifilars you currently have I would expect the top to roll off well into the audio band.
I do not see how the current bifilars that you have will fit into this application, As you mentioned you could re-use the cores and get a new pair of bobbins. And a choke to your needed specs is possible. I'll contact you about pricing.
dave
Great!
Did some adjustment of the irons and also added the somewhat odd tube-models.
Did some adjustment of the irons and also added the somewhat odd tube-models.
- Attachments
-
- LCR6H30CFbalfreq.asc
- (17.16 KiB) Downloaded 637 times
Brgds
Lars
Lars
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Hey Lars,
Looked at your most recent drawing and when I was trying to trace coupling i noticed the parafeed PP coupling. I'm not exactly sure when that popped up but what once was a very simple design is quickly becoming a 600 pound gorilla.
Don't get me wrong, I have very specific beliefs on this which fall well outside the norm. Speaking specifically about the coupling between the input tube and the CF, a pair of bifilars on a common core should get you there clean and simple, Plus bandwidth out to 200K and beyond shouldn't be an issue. Phase match with the bifilars should be WAY better than any other transformer method.
They way i see it, If you are going to use the chokes to load the input tube, Make them bifilar and be done with it and all of the rest of the crap you need to get your signal down the line is ultimately going to give you headaches.
Dave
Looked at your most recent drawing and when I was trying to trace coupling i noticed the parafeed PP coupling. I'm not exactly sure when that popped up but what once was a very simple design is quickly becoming a 600 pound gorilla.
Don't get me wrong, I have very specific beliefs on this which fall well outside the norm. Speaking specifically about the coupling between the input tube and the CF, a pair of bifilars on a common core should get you there clean and simple, Plus bandwidth out to 200K and beyond shouldn't be an issue. Phase match with the bifilars should be WAY better than any other transformer method.
They way i see it, If you are going to use the chokes to load the input tube, Make them bifilar and be done with it and all of the rest of the crap you need to get your signal down the line is ultimately going to give you headaches.
Dave
Dave,
This is almost exactly what I was asking you about from the beginning(thu july 03) but you turned my idea down. Maybe because I called it quadfilar?
Anyway I am also concerned about the low end. The 6F12PT(mu 100, Gm 19mA/V) has an Zout in the ballpark of 5k. So a quite high inductance should be needed.
Will this also be doable on the same core?
The parafeed was actually the the only way to get the response down to 20Hz by careful choice of cap-value.
About the following sentence I don´t understand exactly what you mean in the last phrase:
"They way i see it, If you are going to use the chokes to load the input tube, Make them bifilar and be done with it and all of the rest of the crap you need to get your signal down the line is ultimately going to give you headaches."
To be sure that I got it right: Is this the way you want the driver to CF dual coil bifilar? If not do the adjustments in the file and return it.
This is almost exactly what I was asking you about from the beginning(thu july 03) but you turned my idea down. Maybe because I called it quadfilar?
Anyway I am also concerned about the low end. The 6F12PT(mu 100, Gm 19mA/V) has an Zout in the ballpark of 5k. So a quite high inductance should be needed.
Will this also be doable on the same core?
The parafeed was actually the the only way to get the response down to 20Hz by careful choice of cap-value.
About the following sentence I don´t understand exactly what you mean in the last phrase:
"They way i see it, If you are going to use the chokes to load the input tube, Make them bifilar and be done with it and all of the rest of the crap you need to get your signal down the line is ultimately going to give you headaches."
To be sure that I got it right: Is this the way you want the driver to CF dual coil bifilar? If not do the adjustments in the file and return it.
- Attachments
-
- Daves.asc
- (11.09 KiB) Downloaded 567 times
Brgds
Lars
Lars
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Lars.
That is exactly how i think it should be done, but this cannot be wound quadfilar so that my indeed be where the miscommunication came from.
L9 and L14 should be a bifilar pair and L13 and L15 a bifilar pair and they should be assembled on a common core.
With this design, you should have plenty of inductance to assure low frequency extension and the top end extension and phase relationship will also be better than any discrete sectioned transformer out there.
dave
That is exactly how i think it should be done, but this cannot be wound quadfilar so that my indeed be where the miscommunication came from.
L9 and L14 should be a bifilar pair and L13 and L15 a bifilar pair and they should be assembled on a common core.
With this design, you should have plenty of inductance to assure low frequency extension and the top end extension and phase relationship will also be better than any discrete sectioned transformer out there.
dave
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
hey lars.
Since the basic concept is the same as what John used for his BDT amp we can use his transformers as a starting point. They were ~200R per winding (400R end to end * 2) and had about 520hy's end to end (130 hy per) and can take 1ma of offset.
The cores you have are one size bigger which allows you some options. You can get more inductance, Lower DCR , more possible offset or small improvements in each.
my suggestion is to let the maximum DCR set that point. The DCR will also tell me the number of turns and wire etc to do the bobbins. Then you can play with the gap size to see the tradeoff between sonics, offset limitations, inductance and linearity.
dave
Since the basic concept is the same as what John used for his BDT amp we can use his transformers as a starting point. They were ~200R per winding (400R end to end * 2) and had about 520hy's end to end (130 hy per) and can take 1ma of offset.
The cores you have are one size bigger which allows you some options. You can get more inductance, Lower DCR , more possible offset or small improvements in each.
my suggestion is to let the maximum DCR set that point. The DCR will also tell me the number of turns and wire etc to do the bobbins. Then you can play with the gap size to see the tradeoff between sonics, offset limitations, inductance and linearity.
dave
Whats the penalty of a high DCR? As we are driving into a hiZ load a low secondary DCR shouldn´t be needed.
As the drivers have a Zout in the ballpark of 5k(mu/Gm=100/20) a really high H is needed to lighten the driver load at 20Hz and get a good response down there. Is 200-300H per side doable?
Think the fairly high Gm(20mA/V) russian drivers could be hard to match perfectly so my guess is that an offset of 1-2mA is needed.
As the drivers have a Zout in the ballpark of 5k(mu/Gm=100/20) a really high H is needed to lighten the driver load at 20Hz and get a good response down there. Is 200-300H per side doable?
Think the fairly high Gm(20mA/V) russian drivers could be hard to match perfectly so my guess is that an offset of 1-2mA is needed.
Brgds
Lars
Lars
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
I guess the natural progress is a learning one.
For the first pair of bobbins i used two different wire sizes to improve fill and get as many turns on as possible. The DCR's summed up to about 825 ohms per half and the plots are below. I Had the feeling given the results you would think that going up in DCR would be a good move and the next version used some twisted pair wire i have and came in at 1800 ohms per half.
It is important to note that the plots below are for the plate to B+ winding and the current measured represents what the imbalance would be. For the end to end inductance it is safe to multiply the L scale by 4 and since this is class A the load seen by each tube 2X the L plotted (or 1/2 the end to end)
In any event a pair of both exist (and an even higher DCR option is possible) so lets discuss.
dave
For the first pair of bobbins i used two different wire sizes to improve fill and get as many turns on as possible. The DCR's summed up to about 825 ohms per half and the plots are below. I Had the feeling given the results you would think that going up in DCR would be a good move and the next version used some twisted pair wire i have and came in at 1800 ohms per half.
It is important to note that the plots below are for the plate to B+ winding and the current measured represents what the imbalance would be. For the end to end inductance it is safe to multiply the L scale by 4 and since this is class A the load seen by each tube 2X the L plotted (or 1/2 the end to end)
In any event a pair of both exist (and an even higher DCR option is possible) so lets discuss.
dave
- Attachments
-
- Picture 1.png (54.72 KiB) Viewed 29422 times
-
- Picture 2.png (34.79 KiB) Viewed 29422 times
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
the 6X6 is indeed an option with any EI stamped lamination.
My gut feel would be to go for the 900R bobbins and a butt gap and be at 150 hy's (each tube see 300) for a single stage -1dB point in the 6hz range.
If more inductance is desired the core could be alternately stacked in 3 sections (approximately 12 lams each) for the 12X12 plot. While the plot is less linear, you have to remember that the signals we are talking about are small by comparison. A .1V @ 20hz signal represents 10 gauss so you will not be traversing a major part of the line.
dave
My gut feel would be to go for the 900R bobbins and a butt gap and be at 150 hy's (each tube see 300) for a single stage -1dB point in the 6hz range.
If more inductance is desired the core could be alternately stacked in 3 sections (approximately 12 lams each) for the 12X12 plot. While the plot is less linear, you have to remember that the signals we are talking about are small by comparison. A .1V @ 20hz signal represents 10 gauss so you will not be traversing a major part of the line.
dave
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Hey Dave,
I finally read the whole thread. I recently had ALLNIC 1500 II SE 600Ohm LCR phono on loan. I liked what I heard and as a result I started dusting off my old LCR files. I have Tango EQ-2L inductors which were used in this thread. But before I start playing with them, I want to check the higher impedance LCR inductors.
How are your 7kOhm LCR inductors? You mentioned earlier in this thread that they work great. Nothing was said about them after that. They would for sure make life easier and eliminate one stage of amplification and maybe a step down transformer. My only concern is how to shield the raw inductors. There was quite a bit of hum when I put raw MC step ups into stainless steel cans with permalloy and copper foils inside. However, I think that there are some transformer enclossures that one can now buy from China or Hong Kong. Steel is not as good as thick permalloy but better than stainless.
Regards,
TonyB
I finally read the whole thread. I recently had ALLNIC 1500 II SE 600Ohm LCR phono on loan. I liked what I heard and as a result I started dusting off my old LCR files. I have Tango EQ-2L inductors which were used in this thread. But before I start playing with them, I want to check the higher impedance LCR inductors.
How are your 7kOhm LCR inductors? You mentioned earlier in this thread that they work great. Nothing was said about them after that. They would for sure make life easier and eliminate one stage of amplification and maybe a step down transformer. My only concern is how to shield the raw inductors. There was quite a bit of hum when I put raw MC step ups into stainless steel cans with permalloy and copper foils inside. However, I think that there are some transformer enclossures that one can now buy from China or Hong Kong. Steel is not as good as thick permalloy but better than stainless.
Regards,
TonyB
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
I have been thinking about what Lars and Dave have said about my LCR phono implementation. If you guys recall, I am loading a high Gm pentode with about 606 ohms to drive the 600 ohm LCR network. I chose this because I wanted to terminate the output of the network using leak grid bias. Anyway, a point was made that the 606 ohm plate load in parallel with the 600 ohm network actually has our pentode working into a 300 ohm load, yikes! Although to my ears it sounds pretty darn good I still wonder if an improvement is possible. Call it audiophile syndrome.
Well, after looking at different possibilities, it occurred to me (Another one of those Duh moments!) that one of Dave’s autoformer attenuators would actually be the ticket. I don’t know why I didn’t think about it before now. I still wanted to drive the LCR network with 600 ohms, but wanted a higher working load for my pentode. A little bit more gain wouldn’t be bad either. So, by taking advantage of the impedance rationing of the autoforrmer, we can actually quarter our drive impedance while only reducing voltage gain by half.
So, instead of loading our pentode with 600 ohms, let’s load it with 2.5K ohms. I chose 2.5K because the Rp in parallel works out close to 2400 ohms. In turn, the autoformer will present the LCR network with a driving impedance of 600 ohms. See schematic for details.
The end result on paper looks like a good compromise. The stage gain should go from 14 to 28 and we will still be driving the LCR network with 600 ohms. I can also retain the DC bias on the LCR network to keep the network’s capacitors energized. The effective load driven by the pentode should now be about 1200 ohms (2.5K in parallel with the reflected 2.4K). This should make the pentode much happier. I’m not really concerned about the sonic impact of the DC blocking capacitor for the autoformer. My passive line level crossover experience has shown me that reactive components shunted to ground have much less influence than series components. The AC signal does not have to pass through the capacitor in order to reach the LCR network. Therefore, I really don’t consider the blocking capacitor to be in the “direct� signal path; at least that’s what I keep telling myself.
In summary, we have quadrupled the load seen by the pentode, and doubled the stage gain. Seems like a good move to me. Why mess around with parafeed transformers when you can do it better with an autoformer? Both still require a DC blocking capacitor. In addition, you can continue to ratio down with the autoformer to get different loads for the pentode.
Rgs, JLH
Well, after looking at different possibilities, it occurred to me (Another one of those Duh moments!) that one of Dave’s autoformer attenuators would actually be the ticket. I don’t know why I didn’t think about it before now. I still wanted to drive the LCR network with 600 ohms, but wanted a higher working load for my pentode. A little bit more gain wouldn’t be bad either. So, by taking advantage of the impedance rationing of the autoforrmer, we can actually quarter our drive impedance while only reducing voltage gain by half.
So, instead of loading our pentode with 600 ohms, let’s load it with 2.5K ohms. I chose 2.5K because the Rp in parallel works out close to 2400 ohms. In turn, the autoformer will present the LCR network with a driving impedance of 600 ohms. See schematic for details.
The end result on paper looks like a good compromise. The stage gain should go from 14 to 28 and we will still be driving the LCR network with 600 ohms. I can also retain the DC bias on the LCR network to keep the network’s capacitors energized. The effective load driven by the pentode should now be about 1200 ohms (2.5K in parallel with the reflected 2.4K). This should make the pentode much happier. I’m not really concerned about the sonic impact of the DC blocking capacitor for the autoformer. My passive line level crossover experience has shown me that reactive components shunted to ground have much less influence than series components. The AC signal does not have to pass through the capacitor in order to reach the LCR network. Therefore, I really don’t consider the blocking capacitor to be in the “direct� signal path; at least that’s what I keep telling myself.
In summary, we have quadrupled the load seen by the pentode, and doubled the stage gain. Seems like a good move to me. Why mess around with parafeed transformers when you can do it better with an autoformer? Both still require a DC blocking capacitor. In addition, you can continue to ratio down with the autoformer to get different loads for the pentode.
Rgs, JLH
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
Hey John,
Nicely done! I keep looking at the autoformer and keep trying to sneak it into the plate circuit to get rid of the cap below. Of course this means all of the plate current will traverse it which will definitely be a problem. The easy answer is make it bigger but then you run into all of the problems inherent to any 2:1.
If i remember correctly you are DC coupled to the LCR so there is no need for the blocking cap other than to keep the DC current out of the autoformer. I guess i have offset windings on the brain but why not?
Chances are you are already regulating the filaments so make it a current reg and you are 90% of the way there. The only issue then becomes adjusting it. Lately i have been seriously considering an additional CCS buried (and decoupled) in the plate supply, then matching the currents is easy, repeatable and stable.
dave
Nicely done! I keep looking at the autoformer and keep trying to sneak it into the plate circuit to get rid of the cap below. Of course this means all of the plate current will traverse it which will definitely be a problem. The easy answer is make it bigger but then you run into all of the problems inherent to any 2:1.
If i remember correctly you are DC coupled to the LCR so there is no need for the blocking cap other than to keep the DC current out of the autoformer. I guess i have offset windings on the brain but why not?
Chances are you are already regulating the filaments so make it a current reg and you are 90% of the way there. The only issue then becomes adjusting it. Lately i have been seriously considering an additional CCS buried (and decoupled) in the plate supply, then matching the currents is easy, repeatable and stable.
dave
- Attachments
-
- a crude sketch... made up the currents so the math was easy
- Untitled.jpg (65.7 KiB) Viewed 28764 times
I hear you on trying to sneak the autoformer into the plate circuit. However, wouldn’t that be a problem for pentodes? Wouldn’t the inductance required for decent LF be huge? (i.e. -1dB at 20Hz with Rp = 46K would be like 1420H!) We would be driving the autoformer directly with the plate instead of the parallel combination of plate and plate resistor? I’m I thinking correctly here?
Rgs, JLH
Rgs, JLH
-
- Posts: 2085
- Joined: Sat Jan 22, 2005 3:54 am
- Location: NYC
- Contact:
from the tubes perspective, it doesn't care where the autoformer is it simply appears in parallel with the 2K5 resistor which Is why I'm reasonably sure you can simply parallel both of them in the plate circuit.
I'm guessing you could even place a 600 ohm resistor off the secondary of a 2:1 transformer and end up with the same results. I seem to be on a "primary loading" kick lately which is why i put the load across the primary. I'm not saying its a good idea mind you
dave
I'm guessing you could even place a 600 ohm resistor off the secondary of a 2:1 transformer and end up with the same results. I seem to be on a "primary loading" kick lately which is why i put the load across the primary. I'm not saying its a good idea mind you
dave