@Dorin Bodea, nope! I said standard GOSS or best M6 EI cores. With EI cores, for example, the absolute minimum you need for 15W @30Hz with LOW distortion is the EI-120/50 with 40 mm x 50 mm core section. As the specific weight of EI 120 is approx. 720 g for each cm of laminations column, the core alone will weight 3.6Kg and the finished transformer 5Kg+.
I have done it in a number of variants: Ale Moglia's 5.6K transformer for 814SE, 4.2K for 300B and 11K for 211. If you use a smaller core, you will have to accept high induction which will result in higher distortion. Standard C cores have similar hysteresis/losses but are 100% oriented. So they are bit better than EI but above 1T induction distortion will rocket. Hi-B cores that can really work up to 1.6T induction are at another level in this regard and can be a lot lighter for the same application.
I have done it in a number of variants: Ale Moglia's 5.6K transformer for 814SE, 4.2K for 300B and 11K for 211. If you use a smaller core, you will have to accept high induction which will result in higher distortion. Standard C cores have similar hysteresis/losses but are 100% oriented. So they are bit better than EI but above 1T induction distortion will rocket. Hi-B cores that can really work up to 1.6T induction are at another level in this regard and can be a lot lighter for the same application.
Last edited:
@45 : No steel factory (which Lundahl is not) will sell metal sheet with distortion specs at lower frequencies, as in 20-40Hz. Instead they measure saturation and induction at 50-60Hz, under specific annealing condition. As such maybe about a dozen sorts can qualify as grain oriented silicon steel (GOSS) with saturation levels 1.2-1.9T, where you can find the so called HI-B which, surprise!, stands for "high magnetic induction grain oriented silicon".
They measure magnetization once the core is assembled. That's what one needs to know. I said it will result in lower distortion of the amplifier. What matters is not just the number but how wide is the range for which magnetization is linear to a good degree. EI cores nominally saturate at 1.6-1.8T too but above 0.8-0.9T magnetization starts to bend quite a lot and that will result in higher distortion of the amp if you run the output transformer at induction above 1T. EI cores are only 2/3 grain oriented and that also has an impact on how saturation happens. That's why one needs a bigger size. You can buy the HiB from Nicore or other established Asian manufacturers VERY easily. They are very good.
We have already discussed these things several times, if wanna do it again open a new thread please. Or resume an old thread.
Last edited:
With utmost respect to our colleagues here, especially to the author of the subject (who has the first say when his thread derails or not) I must reply to 45's last entry as I'm not interested in reading or intervening in old threads, I simply don't have enough time for that.
* It will also be my last reply to this matter, I'm sure 45 and his ego will understand 🙂
If you meant what transformer company measures then the result can be different but not by much - see below "BF", also depends on reannealing
Following your logic I will say that a single E+I lamination pair has 3/4 favorable grain orientation paths, but this doesn't mean that in the 1/4 area the reluctance is higher or more destructive than the airgap itself. Reflect on this, it's worth it.
Of course, when the laminations are interleaved and overlap the reluctance fades.
Long story short the magnetism don't work like that; if one uses the same sort of steel, both C-core and E+I having, say, 0.3mm thickness and loss of 1W/kg then you don't need a double weight for E+I to show exactly the same magnetization curve. Maybe 10-30% more but not double, I've tested both in output transformers and interstages - here's a first hand information but, of course, only valid for a certain type of steel in both cases.
More than that, what I'm trying to explain here is called in industry "building factor" BF, which is defined by BF="transformer loss (W/kg)"/"material core loss (W/kg)". Comparing BF of C-core with rectangular laminations with joints at 45 and 90 degrees renders something in the range 1.1-1.3, rarely 1.4.
Yes, "loss" in industry translates somehow tragically as "distortion" in audio 🙂 - in both cases can be compensated by weight/sectional area
Take care! 🙂
* It will also be my last reply to this matter, I'm sure 45 and his ego will understand 🙂
A steel factory uses industrial standard for measuring loss and permeability of electrical steels sheets, which is Epstein frame. They do it under certain annealing conditions, which can be reproduced entirely or partially once the metal sheet is machined into desired xfmr shape.
If you meant what transformer company measures then the result can be different but not by much - see below "BF", also depends on reannealing
Following your logic I will say that a single E+I lamination pair has 3/4 favorable grain orientation paths, but this doesn't mean that in the 1/4 area the reluctance is higher or more destructive than the airgap itself. Reflect on this, it's worth it.
Of course, when the laminations are interleaved and overlap the reluctance fades.
Long story short the magnetism don't work like that; if one uses the same sort of steel, both C-core and E+I having, say, 0.3mm thickness and loss of 1W/kg then you don't need a double weight for E+I to show exactly the same magnetization curve. Maybe 10-30% more but not double, I've tested both in output transformers and interstages - here's a first hand information but, of course, only valid for a certain type of steel in both cases.
More than that, what I'm trying to explain here is called in industry "building factor" BF, which is defined by BF="transformer loss (W/kg)"/"material core loss (W/kg)". Comparing BF of C-core with rectangular laminations with joints at 45 and 90 degrees renders something in the range 1.1-1.3, rarely 1.4.
Yes, "loss" in industry translates somehow tragically as "distortion" in audio 🙂 - in both cases can be compensated by weight/sectional area
Take care! 🙂
Last edited:
OK, so the conclusion is always that with EI cores and standard C cores one has to use a bigger size to get the same result of HiB cores. That's all one needs to know. I still say that 10-30% is not realistic.
Instead of guessing, make a transformer with the specs of the LL1620: 11.5K, 13W at 30Hz and see if you can get the same result in the same application. You are dispensed from the fact that on a 2.5 Kg EI core (the Lundahl is 2.5Kg for the complete transformer) you will have accept significantly higher copper loss to achieve the specs above.
EI cores are 2/3 grain oriented. That is a fact and can be found in literature. This has some negative affect on HOW saturation is approached and happens. Again for the final user, distortion from the output transformer will start to show up at lower induction level than 100% grain oriented. Ranking from low to high is still EI > Standard C > HiB.
Instead of guessing, make a transformer with the specs of the LL1620: 11.5K, 13W at 30Hz and see if you can get the same result in the same application. You are dispensed from the fact that on a 2.5 Kg EI core (the Lundahl is 2.5Kg for the complete transformer) you will have accept significantly higher copper loss to achieve the specs above.
EI cores are 2/3 grain oriented. That is a fact and can be found in literature. This has some negative affect on HOW saturation is approached and happens. Again for the final user, distortion from the output transformer will start to show up at lower induction level than 100% grain oriented. Ranking from low to high is still EI > Standard C > HiB.
Last edited:
@Dorin Bodea you can find it yourself in the net. I was talking about EI core.
You are surely a talker but I don't think you have ever made a 845 output transformer with EI core otherwise you would know that EI120/50 is the MINIMUM to get good performance for nominal 15W @30Hz SE amp, anything smaller would be a significant compromise and you would not be here arguing on pointless things like Epstein, industry and other useless stuff. In addition to the core size you also need space to fit more insulation to work safely at much higher voltage than usual and in this respect C cores are better. That's another valid reason to avoid smaller size.
You are surely a talker but I don't think you have ever made a 845 output transformer with EI core otherwise you would know that EI120/50 is the MINIMUM to get good performance for nominal 15W @30Hz SE amp, anything smaller would be a significant compromise and you would not be here arguing on pointless things like Epstein, industry and other useless stuff. In addition to the core size you also need space to fit more insulation to work safely at much higher voltage than usual and in this respect C cores are better. That's another valid reason to avoid smaller size.
Last edited:
@45 I'm not interested in having a dialogue with you since you've proven you don't understand basic magnetism - not to mention the twisted logic - so from my side you can continue selling snake oil here. Besides, why do you call others names and assume what they did or didn't do something; in your rationale does that invite to dialogue or just escalates some more useless confrontation? I've built xfmrs from MC step-up to GM70 and GU81 output transformers, including interstages and power supply, the whole 9 yards, satisfied? As a "talker", how you called me, I have this page where I put from time time (when I remember) some of my work, if you still have troubles believing me 🙂 - https://www.facebook.com/profile.php?id=100057288107123
@Dorin Bodea you are funny at the very least because you quoted me in the first place but now you don't want to have a discussion because you understand everything and I don't.
You are selling snake oil because you have no idea how to estimate the weight of the core for a given application and use hyperboles to describe anything except what really matters!
Have you ever seen how induction is distributed across the core when grain orientation is not 100%??? Clearly not, otherwise you would not be arguing. It's not uniform and that means that while the AVERAGE induction is still well below nominal saturation there are hot spots where the core is SATURATING. These hotspots eventually will grow until it's fully saturated. In EI cores this can be seen already when AVERAGE induction is around 1.2T. In 100% oriented this does not happen. There are endless articles on this subject both experimental and theoretical with FEA calculations etc.
Epstein measurements in GENERAL are USELESS for explaining EI cores because are always done, as a rule, on 100% oriented laminations and not on the commercial EI laminations. Ask any manufacturer. Actually you can already read that on the Edcor website. And even you used Epstein on partially oriented core, you would still see the AVERAGE with minor difference in magnetization (what I was saying before). But in reality effect is more clear on distortion of the amp at low frequency.
In your fbk link I cannot see any big output transformer. Sorry I am not on fbk. So if you have done it post here. The only output transformer from images I can see is MAYBE one for a 300B? It doesn't look as small as the Lundahl and yet it's for no more than 8W?.
Have you done this 2.5-3 Kg OPT on EI core that performs as the LL1620? I don't think so otherwise you won't be here screaming like a baby. I did not say you don't make ANY transformer, I did say that you have never made a 845 transformer that weights as the Lundahl or max 30% more as you have been claiming since you started your hyperbole....
You are selling snake oil because you have no idea how to estimate the weight of the core for a given application and use hyperboles to describe anything except what really matters!
Have you ever seen how induction is distributed across the core when grain orientation is not 100%??? Clearly not, otherwise you would not be arguing. It's not uniform and that means that while the AVERAGE induction is still well below nominal saturation there are hot spots where the core is SATURATING. These hotspots eventually will grow until it's fully saturated. In EI cores this can be seen already when AVERAGE induction is around 1.2T. In 100% oriented this does not happen. There are endless articles on this subject both experimental and theoretical with FEA calculations etc.
Epstein measurements in GENERAL are USELESS for explaining EI cores because are always done, as a rule, on 100% oriented laminations and not on the commercial EI laminations. Ask any manufacturer. Actually you can already read that on the Edcor website. And even you used Epstein on partially oriented core, you would still see the AVERAGE with minor difference in magnetization (what I was saying before). But in reality effect is more clear on distortion of the amp at low frequency.
In your fbk link I cannot see any big output transformer. Sorry I am not on fbk. So if you have done it post here. The only output transformer from images I can see is MAYBE one for a 300B? It doesn't look as small as the Lundahl and yet it's for no more than 8W?.
Have you done this 2.5-3 Kg OPT on EI core that performs as the LL1620? I don't think so otherwise you won't be here screaming like a baby. I did not say you don't make ANY transformer, I did say that you have never made a 845 transformer that weights as the Lundahl or max 30% more as you have been claiming since you started your hyperbole....
Last edited:
The amplifier is perfect as it is, I feel it, I hear it, I know it. Just wanted to drow some attention on it, you obviously and grotesquely ignored it at the beginning.
If I what to straighten SQ I can always use GNFB, if -6dB is too much, I will use less.
Of course I can go high way and use an output tranfo like ISO specially designed for 845. Let me know if you are interested in seeing how an ISO measures.
If I what to straighten SQ I can always use GNFB, if -6dB is too much, I will use less.
Of course I can go high way and use an output tranfo like ISO specially designed for 845. Let me know if you are interested in seeing how an ISO measures.
Filenet that is not quite true. There were some reservations about your design, which were solved. I asked why you took a scaled down 845 output stage and your reply made sense. When the amp as it is now is able to drive the loudspeakers to your liking (which is where it is about finally) then your project has succeeded.
When you are able to show ISO transformer measurements...that would be interesting.
When you are able to show ISO transformer measurements...that would be interesting.
A 845 amp with 800V supply and 11K load is not a scaled down design. The amp is 12W (including losses) @30Hz without distortion but it is definetly 15W from about 35Hz all the way up. Many amps, including commercial ones, are much worse at 30Hz respect to claimed output power. So, it call certainly be called 15W amp in practical use and is pretty much standard. It's just a different load condition respect to 15W amp in datasheet.
Igor Popovich does it all the time, 2/3 of his designs, are not using max P anode. I wanted to get all the benefits of 845 without hernia as one user say. I like the slam of big transmitting DHT tubes. Also I like the delicacy of small DHT tubes, that's why I choose AD1. 400Vpp is enough for AD1 to imprint his character to 845.
I know that people say that tubes don't have character, only circuits have it. Try and replace VT25 with VT25A in the same circuit.
At high volume it sounds like an 100W SS amp, if this is what you mean.
I know that people say that tubes don't have character, only circuits have it. Try and replace VT25 with VT25A in the same circuit.
At high volume it sounds like an 100W SS amp, if this is what you mean.
Last edited: