Advice on transformer ratios

the road not taken.
Post Reply
montesquieu
Posts: 7
Joined: Sat Mar 07, 2020 9:32 pm

Advice on transformer ratios

Post by montesquieu »

Dave, I wonder if you could spare some advice from your experience of creating SUTs for the Audio Note Io. I have an Io2 on the way (0.05mV output, 1 ohm, the AN UK version).

At some point I reckon I will need to bite the bullet and go for a pair of custom silver wire transformers (a dedicated custom device surely being the best approach) but in the short term I was wondering if the SUT I have would be suitable, and if so what sort of settings I should be looking at?

I have a Miyajima ETR-Stereo. This is an interesting bit of kit, it has four taps on the primary coil (80, 120, 180 and 270 turns, plus pass) and four on the secondary (1000, 2000, 4000 and 8000, plus pass). It also has loading resistors on the secondary, I’m not sure what the values are but the knob is marked 3.6k ohm, 8k ohm, 15k ohm, 24k ohm, 36k ohm and 47k ohm (pass). It also has a knob for capacitance, which is handy for some of Miyajima’s own cartridges (of which have the Madake and three mono cartridges).

Anyway the primary-secondary options would appear to suggest the SUT is capable of 1:100 ratio (80 primary, 8000 secondary) as well as other options for 1:50 and 1:66, all of which I have read as possibly suitable for the Io.

What I’m not sure is how the loading would work in this scenario for a 1 ohm cartridge – Miyajima’s rule of thumb is the following, using the primary coil as the guide:

80 – 2-5 ohm cartridge
120 – 5-12 ohm
180 – 12-30 ohm
270 – 30-100 ohm

The Io doesn’t fit this being 1 ohm – though it has to be worth a shot?

Anyway, I’d appreciate any advice you can offer on how to handle this and the principles behind the choice. The ETR-Stereo while moderately expensive has been a pretty useful tool for adjusting to cope with a sizeable stable of cartridges. Phono stage is an Allnic H7000V, pretty standard 47k MM input.

Another option is to find a way to boost the Io signal into an Allnic HA3000 head amp, I got this for use on the mono side of things (to avoid ground loops from single coil mono cartridges). This has 30db of gain – 1:29 ratio, more or less – and impedance settings of 50, 100, 250 and 500. This obviously is not enough for the Io, but perhaps an option would be a custom SUT to boost the Io signal sufficiently to go into the Allnic – something between 1:4 and 1:7 – though the question again is how the loading would work? Is this a sensible option or a daft idea?

I'm not technical in any sense, I have quite a bit of experience with cartridges and SUTs, but more in a rule of thumb sort of way than through a proper understanding of the maths involved.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Hey,

It seems that you have the swiss army knife of SUT's and my only warning is that just because it can provide a given ration doesn't mean it will be optimized for your setup.

If it does indeed have the ability to do 1:100 and since you already have it give it a try and see what you hear. The biggest problem with a 1:100 turns ratio is the impedance ratio is 1:10,000 so the 47K input resistor reflects back 4.7Ω which will lose you around 2dB of gain from the beginning. Then consider what the cable and input capacitance of the phono do to the signal @ 20Khz and you will be down another 1dB or more.

The above all assumes that given your 1Ω source and 47K || 300pf or so load you do not have any excessive ringing near the audio band.

It also looks like the miyajima can be wired for around a 1:4 that might be worth a try with your head amp but again how the cartridge and transformer behave with each other should really be confirmed before any final judgements on the merits of the SUT concept can be made.

If you try it and like it, that is all that matters. If you try it and don't like it it is best to look deeper before you abandon the step-up path.

dave
Get Your Fix
www.hifiheroin.com
montesquieu
Posts: 7
Joined: Sat Mar 07, 2020 9:32 pm

Post by montesquieu »

Thanks Dave it is indeed a bit of a swiss army knife, I've used it with everything from a 1.5v Ortofon CG25 mono at 1:3.7, to an Ikeda 9TT (2ohm/0.16mV) which was pretty happy at 1:33.

I read your comments on a previous question about the Io, about possibly 1:50 being a more suitable ratio so I'll try that, and 1:66 as well.

Some additonal questions if that's ok :) -

First of all I assume what you mean by that low reflected impedance would be a possible impact on high-frequency performance ... I take it that is what might potentially be improved by using one of the lower ratios above? (Fewer losses and a higher reflected impedance).

Also (and sorry if this is a noddy question), I've read the term 'ringing' before and that it's visible as distortion on a square wave, but what would be its audible effects?

Is there any role for the secondary loading in any of these scenarios?

I will be in touch anyway for a quote on a pair of silver transformers ... hopefully the experimentation will be illuminating in terms of the ratio that works best in this setup with this arm/arm cable and phono stage.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey,
First of all I assume what you mean by that low reflected impedance would be a possible impact on high-frequency performance ... I take it that is what might potentially be improved by using one of the lower ratios above? (Fewer losses and a higher reflected impedance).
when you get to the point where the reflected load = the internal cartridge impedance you will lose 6dB of output. the perfect example of this is a 30Ω cartridge into a 1:20 and a 1:40. Assuming 47K the 1:20 will reflect back 117Ω and the 1:40 will reflect back ~30Ω. 30Ω into 117Ω will net a 2dB loss. 30Ω into 30Ω nets a 6dB loss so you will see that going from 1:20 to 1:40 in this situation which should be a 6dB net gain only gets you an additional 2dB. Granted this is an extreme case but it is only part of the big picture.

The other thing that comes into play with the low Z carts is the Tonearm cable resistance adds to the cartridge impedance from the viewpoint of the SUT and in the case of the 1Ω carts the tonearm wiring could double that value or more. The DCR of the transformer primary also sneaks in and can lose another dB or two so going from 1:50 to 1:100 in your case might only net you 2dB of gain.

The first thing that should be done is get that 47K resistor out of there... it serves no purpose and is a liability for MC cartridges through a SUT.

Also (and sorry if this is a noddy question), I've read the term 'ringing' before and that it's visible as distortion on a square wave, but what would be its audible effects?
Ringing as a generic term doesn't need to necessarily be a bad thing. Ringing near the audio band is bad but if the ringing happens a decade above the audio band it is not necessarily a problem that needs to be dealt with. These plots of the 4722 illustrate that.

Image

The turquoise trace is with a 15Ω source into the 1:18 setting and you can see a 7-8dB peak at 20Khz which gives a very bright forward aggressive sound. Magenta is the same transformer with a 40Ω source into the 1:36 connection and suddenly you are -4dB @ 20Khz which leads to a dull lifeless sound. All of these plots were done without any secondary loading and if the typical 47K load is added the magenta plot would lose 6dB of output matching the two middle plots and the smallish 5dB peak would be damped giving flat response to 20kHz and a rolloff after that. This is why this transformer gained cult like following with the denon 103 and gets called a piece a vintage garbage by anyone who tries it with a 2Ω cartridge.

Is there any role for the secondary loading in any of these scenarios?
If you ask me... loading a transformer is the worst thing you can do to it sonically. It is simply a band-aid to cover up the use of the wrong device. Sure it may sound better with a specific load but it is never representative of what proper design can do. I firmly believe that the people who adjust secondary loads to effect different cartridge loads are often changing the sound of the SUT more than that of the cartridge.
Get Your Fix
www.hifiheroin.com
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

hey,
First of all I assume what you mean by that low reflected impedance would be a possible impact on high-frequency performance ... I take it that is what might potentially be improved by using one of the lower ratios above? (Fewer losses and a higher reflected impedance).
when you get to the point where the reflected load = the internal cartridge impedance you will lose 6dB of output. the perfect example of this is a 30Ω cartridge into a 1:20 and a 1:40. Assuming 47K the 1:20 will reflect back 117Ω and the 1:40 will reflect back ~30Ω. 30Ω into 117Ω will net a 2dB loss. 30Ω into 30Ω nets a 6dB loss so you will see that going from 1:20 to 1:40 in this situation which should be a 6dB net gain only gets you an additional 2dB. Granted this is an extreme case but it is only part of the big picture.

The other thing that comes into play with the low Z carts is the Tonearm cable resistance adds to the cartridge impedance from the viewpoint of the SUT and in the case of the 1Ω carts the tonearm wiring could double that value or more. The DCR of the transformer primary also sneaks in and can lose another dB or two so going from 1:50 to 1:100 in your case might only net you 2dB of gain.

The first thing that should be done is get that 47K resistor out of there... it serves no purpose and is a liability for MC cartridges through a SUT.

Also (and sorry if this is a noddy question), I've read the term 'ringing' before and that it's visible as distortion on a square wave, but what would be its audible effects?
Ringing as a generic term doesn't need to necessarily be a bad thing. Ringing near the audio band is bad but if the ringing happens a decade above the audio band it is not necessarily a problem that needs to be dealt with. These plots of the 4722 illustrate that.

Image

The turquoise trace is with a 15Ω source into the 1:18 setting and you can see a 7-8dB peak at 20Khz which gives a very bright forward aggressive sound. Magenta is the same transformer with a 40Ω source into the 1:36 connection and suddenly you are -4dB @ 20Khz which leads to a dull lifeless sound. All of these plots were done without any secondary loading and if the typical 47K load is added the magenta plot would lose 6dB of output matching the two middle plots and the smallish 5dB peak would be damped giving flat response to 20kHz and a rolloff after that. This is why this transformer gained cult like following with the denon 103 and gets called a piece a vintage garbage by anyone who tries it with a 2Ω cartridge.

Is there any role for the secondary loading in any of these scenarios?
If you ask me... loading a transformer is the worst thing you can do to it sonically. It is simply a band-aid to cover up the use of the wrong device. Sure it may sound better with a specific load but it is never representative of what proper design can do. I firmly believe that the people who adjust secondary loads to effect different cartridge loads are often changing the sound of the SUT more than that of the cartridge.
Get Your Fix
www.hifiheroin.com
montesquieu
Posts: 7
Joined: Sat Mar 07, 2020 9:32 pm

Post by montesquieu »

Thanks Dave very informative. I will have a play with the various ratios once the Io arrives and be in touch directly.
montesquieu
Posts: 7
Joined: Sat Mar 07, 2020 9:32 pm

Post by montesquieu »

Just a brief postscript to this, the Io2 arrived and works rather well with the Miyajima ETR-Stereo at 1:100 and 1:66 both without any secondary loading required.

1:100 preferable and losses seem to be minimal - there's quite an audible jump in volume between the two settings and 1:100 does seem quite 'right'.

Will still be interesting to pursue a dedicated SUT but that can be done at leisure.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

Do you have an estimated value for your cable capacitance and input capacitance of your allnic? These numbers become substantial when multiplied by the impedance ratio of a 1:100 (10,000)

dave
Get Your Fix
www.hifiheroin.com
montesquieu
Posts: 7
Joined: Sat Mar 07, 2020 9:32 pm

Post by montesquieu »

dave slagle wrote:Do you have an estimated value for your cable capacitance and input capacitance of your allnic? These numbers become substantial when multiplied by the impedance ratio of a 1:100 (10,000)
The Ikeda arm cable specs (1m cable) are:

Conductor Resistance : 32mΩ/m
Capacitance : 52pF/m

My Furutech interconnect between Allnic and SUT (also 1m) is 64pF/m.

I can't find any info about the capacitance requirements of my Allnic H7000V (nor for the very similar H3000V which pre-dated it) but it is an LCR design and all transformer-coupled.
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

for the capacitance:

the e810F as a triode has about 100pf of input capacitance so that puts you in the 225pf range which translates to the cartridge seeing 2.25µf of capacitance. At 20kHz that 2.25µ has an impedance of 3.6Ω (7.2Ω @ 10kHz)

For the reflected load a 100K input resistor becomes 10Ω and a 47K becomes 4.7Ω which appears in parallel with the 3.6Ω so that 100K value is helpful. The 1Ω impedance of the Io keeps this situation from being impossible but the hidden factor that can pop out of the weeds is the tonearm cable DCR. This only becomes an issue with the low Z carts being heavily loaded which is what a 1:100 does.

dave
Get Your Fix
www.hifiheroin.com
Post Reply