[Andrew]

Hi Dave,

What were your conclusions on MC-Step ups in the end?

I have a pair, not yours I might add, that are set to 1:10. I'm using a Denon 103R loaded with 10K to give 100R, and I'm finding that, when compared to a simple solid state head amp, also set to gain of 10 with a 100R load, that they are a bit slow and lacking in the freq extremes, through the midband is much better. The better half descriibed them as muddy by comparison, not my choice of words, but an interesting observation, nonetheless.

cheers,

-- Andrew

[fred76]

Hi dave,

Any updates on the step-ups? Hopefully you can come up with a design for low to very low output/low resistance MC's. Maybe can be configured for either 1:20 or 1:40.

regards,
fred

[dave slagle]

[dave slagle]

[Andrew]

[dave slagle]

[fred76]

[fred76]

Andrew,

IMHO the 103R sounds much better with a higher ratio than 1:10. I personally would go for 1:20 (or higher if needed) so it would be closer to the 100ohm load without loading R's. Some 'forward sounding' systems even require lower loading with the 103R to sound a bit more balanced. It's a "system thing", sometimes you need to compensate for changes w/in the system chain WRT input sensitivies and speaker efficiency etc.

I used to own Sowter 6495, 1:10. It was designed for medium to low resistance carts. It has good response at the LF but really lacks 'openness' even when unloaded. It sounds a bit better with medium resistance carts ranging from .5mV-.8mV. But in the end they still sound pretty hohum especially in the midband.

[Andrew]

Hi Fred,

Just to confirm...

Would you suggest no load at all then? Neither primary nor secondary? And setting for 1:20. How would that work? Just the resistance of the TX secondary itself? I had these set to 1:20 originally and used a standard MM load of 47k which would have provided about 117R at the cart.

thanks,

-- Andrew

[Andrew]

-- Dave,

A couple of questions for you, if I may.

Take a well implemented but reasonable price point example of step up tx.

How good can we expect the real life coupling efficiency to be between its primary and the secondary winding?

How much parasitic parallel capacticance will there be across each winding?

Ball park figures are fine, just trying to get my head round the maths here.

many thanks,

-- Andrew

[dave slagle]

[dave slagle]

[Andrew]

Hi Dave,

Primaries 3 secondaries 0

I put 82R in series with my primaries and unloaded the secondaries so its just the grid to ground resistance of the first value > 1Meg in parallel with the 6R of the primary DCR plus 82R, essentailly 88R.

This sounds much much faster, though I think I prefer a 100R-sh load on the 103. I wil compare against my solid state reference later or tomorrow to definitive answer.

I have a theory and have been playing with some models whilst listening, howeever a lot depends upon getting a reasonable set of paremeters for an "average" step up. Do you think in the interests of science we could work the parameters for one out?

cheers,

-- Andrew

[fred76]

[fred76]

[dave slagle]

[Andrew]

Hi Dave,

Yes scrap this I screwed up, I should have done parallel, but for some reason my brain went haywire, sometime about 10pm last night after I had spent an age stripping my phono down to remove the secondary load resistors.

I started soldering for parallel then measured, got 6R (primary DCR) and thought hmm not enough resistance, better add these resistors in series. What was I thinking? I had forgotten now it was "on the bench" my secondary was disconnected from my load. DOH! Double DOH! So no reflected load.....to be in parallel with when I measured with the DVM (though I'm not even sure what that would have told me)

But now I have my workings all exposed its a 15 minute job to move them to parallel, which is what I intend to do first chance tonight.

Funny thing is though, the speed and attack was back! It sounded quite bright/zippy though, I only had a really quick listen. which in retrospect was probably correct. I think must have ended up with about 10k reflected load onto the cart, basically grid resistance/100. However, the speed was definitely back. Like the secondary load had been causing some sort of HF roll off that I couldn't hear but was enough to cause slew rate limiting.

thanks,

-- Andrew

[dave slagle]

hey andrew,

it is those kind of mistakes that really teach us something

the speed and attack you heard makes me wonder what a slightly inductive load might do... sure at some point the iinductance will mess with the FR in the audio band, but we can cotrol that.

see new thread.

[Andrew]

Rematch...

100R on the primary in parallel,

primaries 3 : secondaries 0

100R on the primary definitely sounds better than 10K on the secondary. I don't even need to go back and check. My solid state reference still need to be checked, but, the speed is there, so its the imaging, getting this right got rid of the fizz and brightness caused by the large load.

Now we just need to work out why.

Dave, I have a theory..............

cheers,

-- Andrew

[Johnny]

Hi Andrew,

I had a Denon 103R and really liked it. I found that loading the cartridge with a 70 ohm load gave me the best performance. I tried values from 42 ohms to 120 ohms. When I used higher values the sound was a little hard and tizzy. When I used the lower values, the sound was too soft and dull. The ratio of the load seen by the cartridge and the internal impedance was 5:1 when I used the Denon 103R (i.e. 70/14). I've found with other cartridges that this ratio can vary from 3:1 to 9:1. You might want to try values around 100 ohms.

Johnny

[Andrew]

Thanks Johnny,

Funnily enough, I'm using 100R. I found 82R a bit dull in my system and 100R brought up the treble nicely. Loading the primary, of course

The sound is much better than with a secondary load. Later I hope to compare against my little solid state head amp that cost me all of 10$ to build and started me off down the road of realising my step ups weren't doing what they should.

cheers,

-- Andrew

[IslandPink]

Andrew ,
I spoke to James and he mentioned your experience with putting the load on the primary, instead of the secondary, with the S&Bs . It hadn't occurred to me that this option might be better . Tried it , replacing 10K on the secondary with 120R on the primary ( and 100k on the secondary ie. grid-leak for the D3a ) and ....

( i) First I thought it was just brighter
(ii) Then I thought it was brighter but also better with more dynamics and tone in the treble .
(iii) The next day I thought it sounded worse, too bright , and less tone in the midrange .

Ok, so now I have to go back to loading on the secondary , because I reckon there are pros and cons . Theoretically, I'd expect loading the secondary was better , to get a bit more current flowing in the transformers .

I'll let you know when the dust has settled on this mod .

Mark

[Andrew]

Hi Mark,

I have to say, whilst I understand your point about then current, this one's a done deal as far as I'm concerned. How about trying loading the primary but with a higher load, to bring down the zip and ting a bit? I would have expected it to sound like 2) in your report.

I have a theory on this one and Spice confirmed it, now I'm trying to pin down what Spice does so that I know its behaviour is "appropriate" for the model. I have to say I don't doubt it is, but I'd like text book theory to coincide with the computer model.

cheers,

-- Andrew

[nickg]

Just FYI, I have tried your suggestion, loading the primary (lundahl step ups), it certainly made a difference, but I have gone back for the moment, I will try again later. It seemed to my ears to bring up the lower midrange, also there was a lot more lower bass, and the top end was more detailed as well (I know that means everything got louder, but it wasn't just that). Whilst all that sounds like it should be good, the reason I have gone back, is there was a feeling of much more information, but it didn't seem to make as much sense, the music wasn't as coherent, more a sequence of hifi moments. Maybe thats why all the parts seemed to be emphasised, with the cohesion gone, all the parts were able to shout on their own, instead of being marshaled into a whole.

Not sure if the above makes sense, but hopefully you can pull some meaniong from it. Andrew, I have all the bits to do that SS stepup, I am saving it for a booring bit of the holidays before building. We will see how that compares.

[Andrew]

Well, well, this is really very very intruiging! A puzzle indeed...

My niave assumption was that this was a "no brainer" but obviously its not in the slightest, at least, in some systems, or to some folks ears/listening tastes.

Nick, thanks, your info is really worthwhile as you have a similar cart but different step ups. Please do post when you've built the SS jobbie, that will make a very useful comparison. I find my system really slow when loaded on the secondary now....I think Greg might have inadvertantly hit on this at the phono shootout as well as he kept complaing whenever the S&Bs were in the loop.

Mark's data point is has me intrigued as well.......

Dave any thoughts, especially in the light of the Spice sims sent your way that show the roll off...as the efficiency of the primary-secondary coupling decreases...I must mail those to Nick and see what he thinks....Mark are you able to run LTSpice?

How about loading both sides, say 30-70, 50-50 and 70-30 one of these might get us a nice mix?

cheers,

-- Andrew

[dave slagle]

hey guys,

one observation that needs to be made is when you load the secondary you are loading both the cartridge and the transformer so i would expect different trannies to behave differently in this respect.

I generally try to avoid the need for external loads to get stuff to behave, but in other cases moving the loading from secondary to primary might uncover some unwanted step up behavior. The load on the primary also lowers the source Z driving the trannie slightly so that to could also emphasize some behavior of the trannie.

a simple frequency sweep of the two situation might tell us something. If you do this be sure to use a source Z equal to that of your cartridge. A simple voltage divider will do this. Pick a resistor around the desired source Z and then pick a value say 50X greater to form a 50:1 divider. This will give the desired source Z when driven from a 50R generator and it will attenuate the signal sufficiently so we are closer to the levels involved.

dave

[nickg]

It may be adding too much to the complexity, but it could have bearing as to why its not the clear success Andrew found. I have been thinking about the input of the phono stage, but from the perspective of the first grid. Andrew, you know that when I built the TL LCR stage as T's original design, it didn't work well, I had huge imbalance in gain, and way off FR. It turned out this was because T's circuit put the input E810F into grid current land, and the Lundahl step ups I am using just collapsed and went into melt down with just a trace of grid current, so I rebiased the input stage to get away from that.

But I am wondering, not having the load resistor on the secondary means the only way grid current can get away is via the TX secondary. I wonder if the lack of coherence I was hearing was the transformer impedance causing peaks of grid current from clearing away before the next peak comes along.

T's original circuit worked fine as was, with your step ups, I wonder if this is because they have more ability to deal with grid current.

Just a thought....

[dave slagle]

[Andrew]

I should add, for completeness, that I completely removed the secondary load resistor, so no grid leak resistance either. I'm not sure what the lack of the grid leak will do but to the valve, I'll get my books out, I suppose the grid will be references to gnd through the TX secondary. Anyway, in my case there's no external secondary load, on the TX, at all.

Nick, did you use cathode resistors to bias T's mk1 circuit, in the end? I know why he did this but I would be tempted to try more convential bias, even if it does reduce the standing current a bit.

I will post the sims later.

cheers,

-- Andrew

[Andrew]

-- Gents,

Here's an image grabbed from Spice, what I was trying to show is how as the efficiency of the TX primary-secondary coupling drops, the effect of the secondary 10K loading is to create a low pass filter. The efficiency only has to drop to 99% to get a roll of starting at 10k. However, this doesn't happen so readily on the primary load, all other things being the same. Morgan Jones in his 3rd ed (page 244, Fig 4.21) says at high frequencies a step up, when parasitics are considered, starts to look like a low pass filter.

If you reduce the Zout of the cart the effect is less pronounced and so on. This might be why some folks think the 103 and family are dull, as its got a highish Zout, but that's purely speculation.

cheers,

-- Andrew