How to make an 600 ohm LCR have Zin=1200ohm

the road not taken.
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hi,
Interesting to see that the 600 and the 1,5k do not have the same impedance characteristics.

Wonder why? But it can be due to imperfections and differences in our Spiceprograms.

Brgds
Lars
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Hi All,

Well one pssoble reason is that the 1k5 is not actually quite 1k5, but a bit under.

L1 Input N001 4.7075 Rser=150 Rpar=1e120 Cpar=9e-120
L2 N001 Output 0.10875 Rser=35 Rpar=1e120 Cpar=9e-120
R5 N005 Ground 170k
R4 N004 Ground 35k
C1 N004 Ground 2µ
C2 N005 Ground 50nF
R3 N002 N004 291
R1-1 Input N002 1240
R1-2 N002 N001 1240
R2-2 N003 Output 1k41
R2-1 N001 N003 1k41
R6 N003 N005 175R
.backanno
.end

I'm currently using 100R for R6, moves the mythical 50kHz point down a bit.

cheers,

-- Andrew
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Andrew,

you can post .asc files here for us LT users.

Lars,

what spice are you using?

dave
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Andrew,

you can post .asc files here for us LT users.

Lars,

what spice are you using?

dave
Sure.

I didn't cos Lars doesn't use LTSpice and an image dump only tells half the story so a netlist seemed a better bet.

cheers,

-- Andrew
Attachments
LCR-RIAA.asc
(2.12 KiB) Downloaded 797 times
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hi again,
I use Designsoft Tina and unfortunately I can not import .asc files. Anyway I think I have found out why things differ. The Tangos have much lower RDC: 26 resp. 5ohm relative to the 2,5x-factor that where used initially.

Brgds
Lars
Brgds
Lars
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

This is what came out after beginning to work with LTSpice. Direct-coupling af all the tubes! D3a might be changed to E86C that I have in stock. I am a little suspicious about the semiconductors but I will listen to find out.

I am at the moment soldering the testversion of the cards, see pics below.

The "recovery" amp has (not so very)funny C1 and C7 values but it might be the transformer model. Do not know what is worst: A coupling-capacitor from the LCR to the grid of the 6C45 or the proposed cathode capacitor. In the schematic C7 should be connected from cathode to ground not to B+.

Dave, do you have Spicevalues for a proper 2:1 model? I am also thinking of a choke between 100 and 200H.

If someone wants to sim the circuit I can add the *.asc file and netlist.

At the moment I have borrowed a single LL1676 and a 100H choke from a friend just to see if it works IRL like I want to.
Attachments
LTRIAA.GIF
LTRIAA.GIF (16.34 KiB) Viewed 30031 times
LTRIAACURVE.GIF
LTRIAACURVE.GIF (19.83 KiB) Viewed 30031 times
RIAADRIVER.JPG
RIAADRIVER.JPG (116.84 KiB) Viewed 30072 times
RIAAUT.JPG
RIAAUT.JPG (97.68 KiB) Viewed 30072 times
Brgds
Lars
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey lars,

It is nice how using the source to "load" the filter makes the direct coupling easy. My approach has ben to let a little DC "sneak" through the LCR.

I hae not found transformer models to be very reliable in spice beyond the first order approximations. Where do you plan on using the 2:1, if it is after the 6C45, i'd strongly suggest simply using a 1:1 bifilar in that spot. the 1500 ohm source Z of the 6C45 is plenty to deal with most cabling and i just see the 2:1 forcing you to parafeed and use a more compromised piece of iron.

as always, a .asc is nice to have if you want to upload it.

also if you play with the values of c2 & c3 you should be able to flatten out those (small) peaks at 200hz and 10K (i'm guessing those are in a can and must remain fixed though)

dave
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Dave,
I totally agree, transformers seems to be troublesome to sim. But by knowing the inductance you can anyway get a rough picture of the low end roll-off. Just noticed that the end of the schematic with the output 2:1 tranny was outside the picture.

About your bifilars: What will the approximate inductance be with an Ia of 10-20mA? Fooled around with a 1:1/150H yesterday but could not get it right. Also tried Jacks Ultrapath version with no luck.

The Tango EQ2L just contains the inductors so I can play around! This RIAA correction is the best you can get without parallelling the capacitances to ground with resistors, but then you have DC through the inductors :cry:. If you check my first and second post in this thread, I have a solution to that little problem.

Will come back with another schematic (and *.asc) with better correction and a 1:1 when I come home tonight. Maybe you could PM me Spice figures for the proposed bifilar until then?
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Hi Lars,

Yes, please post the ASC.

thanks,

-- Andrew
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hi lars,

for the 1:1 bifilar 150hy's should be doable with 10-20ma.. see attached plots. The DCR will be in the 1100 ohm per winding ballpark and to sim it i'd just couple an inductor to L1 with the same specs.
Attachments
Picture 2.png
Picture 2.png (34.55 KiB) Viewed 30041 times
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Hi Dave,

Why not a step down? Say by 3? Help drive the line better?

cheers,

-- Andrew
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

OK boys, this is the best I could do: Very near +/-0.09dB 20-20kHz.

Did not know so I went for 0.9999 for the bifilar.

Hope the *.asc works, have been messing around with my dmtriodep.inc a little.

Note my reversed RIAA. It is, as far as I understand, theoretically optimal with a gain of 1x.
Attachments
REVRIAA.JPG
REVRIAA.JPG (58.53 KiB) Viewed 30032 times
LTRIAACURVE.GIF
LTRIAACURVE.GIF (20.86 KiB) Viewed 29991 times
reVRIAAwhitebifi.asc
(7.08 KiB) Downloaded 614 times
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Hi Lars,

RIAA looks pretty good to me, I haven't been able to better +-100mB, see the LCR phono thread with my actual measurements in. Who did the PCBs?

cheers,

-- Andrew
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Andrew wrote:Hi Lars,

Who did the PCBs?

cheers,

-- Andrew
I had the cards made -89 for another cascoded-RIAA that did not sell so well. A friend of mine has a company for PCB-layouts so for him it was an easy task. I can now use the cards for RIAA, Line-amp and power-amp driver sothey are universal.

Took the RIAA right out of the book from Tango with tha addition of the 50kHz that I simmed.
Brgds
Lars
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Dave,
Zout will be Ri 2k + DCR 1,1k = ca 3kohm. Do you think it is low enough? Or could one make a really good bifilar 1:1, 100H with lower DCR?
Brgds
Lars
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

lower DCR is possible but like everything else it comes at a cost. The easiest way to drop the DCR is to lessen the inductance. With an Rp of 2K i would consider looking at around 60hy's of inductance which would put your -1dB in the 10hz range.

if you really need the inductance, then you have one other choice and that is a bigger core. This simply adds cost. The cores i was showing you were 49% nickel and that represents the largest you will get in nickel. Amorphous goes bigger but there is the cost issue. Steel is the other option that will give you the lower DCR at a lower cost, but then you have to listen to steel.

dave
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Is 3k too high? Possibly....

I would say it depends upon what you are driving. I have a magentic volume control made by Stevens & Billington and it really needs a source that has a Zout of less than 1k, less than 200R is even better, otherwise you hear the music looses its pace, rhythm and timing.

That's why I asked Dave about a step down, as a niave choice it would seem ideal.

cheers,

-- Andrew
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

dave slagle wrote: lower DCR is possible but like everything else it comes at a cost. dave
Have measured Gm of the 6C45 in my Avo and the assumption of a Ri in the ballpark of 2k seems to be completely right with the choosen working point. Will have problems even with a 2:1 to come down to 600ohms.

Next step in the chain is a Silk TVC(copy of S&B TX-102) with 700H and 380ohm. I also have a switched, parallelled, primary, resistive loading to make the source see a constant Zin of 30k at frequencies over ca 500Hz and below that falling to 24k at 20Hz.

Got the impression that a TVC would work best with a lower source Zout ? Checked S&Bs site and they said a source of 5k would give 1dB drop at 20Hz.
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Hi Lars,

Yes, I agree, for a TVC a much lower Zout seems to be required I can hear a very real difference in the pace of the music between my existing phono (Zout of about 2k) and a similar source driven by a buffer stage (Zout of about 175R)

cheers,

-- Andrew
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Hi Andrew,
You forgot to tell which one sounded more natural :wink: ! Or do they just sound different?
Just joking, I understand the buffered is better. Are the RIAAs identical except for the buffer?
Brgds
Lars
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

dave slagle wrote: Amorphous goes bigger but there is the cost issue. dave
Is amorphous good or bad with reference to nickel? And how low RDC can you get with that?
Asked about bifilar 2:1 in the Transformer-thread so we leave that discussion there.
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

I've also had a friend's LCR and run it buffered and unbuffered into the TVC, buffered had much better pace, rhythm and timing.

cheers,

-- Andrew
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Andrew wrote: I've also had a friend's LCR and run it buffered and unbuffered into the TVC
cheers,

-- Andrew
Great, thanks! Could you show a schematic of the buffer?
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

It wasn't mine, but it was basically an Anode Follower set to have unity gain; it had a Zout of around 170R. Zout of the LCR on its own was around 2k - 3k, perhaps a bit more.

cheers,

-- Andrew
Last edited by Andrew on Wed Oct 17, 2007 4:16 pm, edited 1 time in total.
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Thanks Andrew!
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Dave, what's the smallest step down you think will work and sound good?

Lars, seems, to me, if you don't have the gain to do a larger step down then perhaps you could consider some sort of buffer on the output and use a 1:1 to avoid the cap; those S&B's need a good driver to come to life.

I'm sort of suggesting separating the gain stage from the line driver?

I use a high gm driver tube for my 300B, the gain stage varies on my mood. My 1:1 interstage is placed in the cathode leg of the tube with a the large cathode follower resistor underneath that, works quite well.

cheers,

-- Andrew
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Andrew wrote: some sort of buffer on the output and use a 1:1 to avoid the cap
cheers,

-- Andrew
It is the DCR of the 1:1 that is the problem at the moment. I am waiting for Dave to present a version with lower DCR. If we do a CF or AF we will get an Zout of maybe 100ohm but the DCR of the 1:1 is still over 1k and is added to the buffers Zout. So if it was manageable to get a DCR of 500 intstead it might be a good match.
Brgds
Lars
reVintage
Posts: 365
Joined: Fri Sep 07, 2007 9:54 am
Location: Sweden

Post by reVintage »

Had to do a buffered version. 38dB gain, Zout RDC+100. Skipped the anode follower as it has serial resistors.
Attachments
reVRIAAwhitebifi2.asc
(7.9 KiB) Downloaded 598 times
Brgds
Lars
Andrew
Posts: 201
Joined: Wed Apr 26, 2006 8:06 am

Post by Andrew »

Hi Lars,

Yes, that's what I meant, topologically speaking, it looks good to me! I have found this to be a good arrangement.

I would favour a slighly higher gain valve then use the 6N30 to drive the TX, but that's a personal choice really as I prefer slightly more than 38db.

Be good to hear Dave's thoughts on the step downs still, however.

cheers,

-- Andrew
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

i'll try to cover everything at once.

I know of some people using my autoformers with a 5687 and a 2.2u output cap with no reported problems in the bass or slowness to the sound. when they replaced the plate choke on the 5687 with a bifilar 1:1 (in the 600R range) a noted improvement was heard. I am partially convinced this is because the reactive CL resonance has been removed. I will add that these same people had also tried a TVC with less than stellar results. As an interesting aside, they refused to try my AVC's based on their TVC experiences and it wasn't until i forced a pair in that the difference was heard. For some reason the TVC's in their setup just sounded broken.

I'll also admit that all the talk of late of buffers and magnetic volume controls really has me perplexed. There are units out there with buffers before and others with buffers after and i have to admit that I just don't get it. My universal experience has been that buffers anywhere near the autoformers simply degrade the sound. Maybe i just prefer a different sound and a different choice of compromise. :-)


WRT the 2:1's I can only offer up one outside anecdote apart from my personal experiences. At one point the guys in colorado were using 2:1's to parafeed couple their RS241's to their 75TL's operating under the premise that the 5K Rp of the 241 wouldn't have a chance against the grid of the 75TL. My distaste for parafeed and small step-down transformers in general forced me to send them some nickel 1:1 bifilars to use as a plate choke. When that improved the sound, I had them eliminate the cap for a 1:1 feeding a 2:1 and then the 2:1 hit the curb and the amp simple became a RS241 1:1 75TL amp.

From a measurement perspective, I do not like iron that shows erratic high Q responses above the audio band and this is typically what you see happening in the low stepdown ratio IT's. The two simple solutions to this is to either use loading to fix the problem or to make the trannie small and parafeed it. I have never liked the results of either of these approaches and hence my pushing in the direction of the 1:1 bifilar.

speaking of the bifilar. In this kind of low level situation, i would sell the farm inductance-wise to use nickel. I could easily get you down to 200 ohms and net you plenty of inductance at the cost of a 12 pound amorphous unit. Something much more reasonable could be had with steel C-cores and i do have some nice 2 mil units that might work, but still my gut says go for the nickel. the attached graph is a lower DCR version of the 49% nickel. do note that the primary and secondary have different values and for Zout you need to add them (this means that the 1K units add 2K to the source Z)

dave
Attachments
this was used to replace a 200+ HY plate choke loading a 5687 cap coupled to an autoformer in a phono stage with good results.
this was used to replace a 200+ HY plate choke loading a 5687 cap coupled to an autoformer in a phono stage with good results.
Picture 6.png (33.04 KiB) Viewed 30238 times
Post Reply