Hi.
As I´ve mentioned earlier on this forum I´m collecting parts for a 813 SE project.
The choise of OPT was (is) a big question here, since this will be a big and very expensive project and I don´t want any junk in that important position.
Finally I decided to use Hammond 1628SE´s, but since they seem to have some serious problems in the upper frequency range they´re out of question.
(See the thread "What´s wrong with Hammond 1628SE?")
Sad, since 813 would give more than 25 fairly linear watts with into that load if they´re pushed a bit into class A2.
The second most logical choice would be a pair of Lundahl C-core OPT´s. The problem here is the low primary inductance, 23H for a 5,6k unit is a bit light I think.
The question is: How much would I suffer from the low inductance?
The plate resistance of a triode wired 813 is about 1,8k and my coarse calculations makes me believe that I´d need about twice the inductance to get full frequency response.
813 builders out there, what are you using?
As I´ve mentioned earlier on this forum I´m collecting parts for a 813 SE project.
The choise of OPT was (is) a big question here, since this will be a big and very expensive project and I don´t want any junk in that important position.
Finally I decided to use Hammond 1628SE´s, but since they seem to have some serious problems in the upper frequency range they´re out of question.
(See the thread "What´s wrong with Hammond 1628SE?")
Sad, since 813 would give more than 25 fairly linear watts with into that load if they´re pushed a bit into class A2.
The second most logical choice would be a pair of Lundahl C-core OPT´s. The problem here is the low primary inductance, 23H for a 5,6k unit is a bit light I think.
The question is: How much would I suffer from the low inductance?
The plate resistance of a triode wired 813 is about 1,8k and my coarse calculations makes me believe that I´d need about twice the inductance to get full frequency response.
813 builders out there, what are you using?
Lundahl 1688
Hi,
You might want to look into the new Lundahl 1688 designed for the 845. 9.2k or 5.5k with 70H at 70mA. The datasheet is on their website.
Michael
Hi,
You might want to look into the new Lundahl 1688 designed for the 845. 9.2k or 5.5k with 70H at 70mA. The datasheet is on their website.
Michael
Thank you, but I found it when I had a closer look at their homepage.
I´ll mail Lundahl and ask if they can make a pair with slightly bigger airgaps, since I´ve found a sweet point at 900V 80mA that gives almost 27W (in light A2 operation).
Going from 70 to 80 ma shouldn´t dump the primary inductance too much, don´t you think? From 70 to 50-60H I´d guess.
BTW: Thanks again for pointing me towards this transformer, I had no idea they had it.
I´ll mail Lundahl and ask if they can make a pair with slightly bigger airgaps, since I´ve found a sweet point at 900V 80mA that gives almost 27W (in light A2 operation).
Going from 70 to 80 ma shouldn´t dump the primary inductance too much, don´t you think? From 70 to 50-60H I´d guess.
BTW: Thanks again for pointing me towards this transformer, I had no idea they had it.
I just posted a mail to Lundahl, will check for an answer tomorrow.
One thing that confuses me is the connections of the secondaries.
The transformer is wound as follows: 50+50: 1+2+2+1+1+2+2+1
To get 9,2k:8 we want a ~34:1 ratio, 100:3 would be good enough I think.
I would do it this way: All "2" windings connected in parallel and all "1" windings in parallel, then those two banks of windings in series = 100:3
Accordning to the Lundahl datasheet some sec. windings shall be connected in a very strange configuration, some of them in antiphase. Why?
One thing that confuses me is the connections of the secondaries.
The transformer is wound as follows: 50+50: 1+2+2+1+1+2+2+1
To get 9,2k:8 we want a ~34:1 ratio, 100:3 would be good enough I think.
I would do it this way: All "2" windings connected in parallel and all "1" windings in parallel, then those two banks of windings in series = 100:3
Accordning to the Lundahl datasheet some sec. windings shall be connected in a very strange configuration, some of them in antiphase. Why?
23H is a bit low.
Yes, 23H is a bit low. XL will equal the reflected load at 39 Hz. This will make the load line both elliptical and lower than 5.6Kohm.
If the 23 H is measured at very low voltage drive, expect it to increase a factor of 2 or 3 as the voltage on the primary increases. If this is the inductance called out in the transformer catalog, it usually won't be the inductance at low drive, but it will be the inductance at a higher drive level. Unfortunately, marketing usually spec's the inductance at it's maximum level (at a high drive level.)
8H per kohm of reflected load is a rule I've heard more than once (or 45H for 5.6K.) 8H/kohm gives us a -3 dB point of 20 Hz. You can correct for low primary inductance a little by adding more bias current. Add enough current so that the load line biased up for a 2.6K load.
2.46K = 1/(1/Xl)^2 + (1/RL)^2 = 1/[1/(2*PI*20*23H)^2 + 1/5.6k^2]
Besides the primary inductance, keep an eye on the leakage inductance and total parasitic capacitance on the primary side. Usually, not always, the secondary side's capacitance is low enough not to be an issue.
Yes, 23H is a bit low. XL will equal the reflected load at 39 Hz. This will make the load line both elliptical and lower than 5.6Kohm.
If the 23 H is measured at very low voltage drive, expect it to increase a factor of 2 or 3 as the voltage on the primary increases. If this is the inductance called out in the transformer catalog, it usually won't be the inductance at low drive, but it will be the inductance at a higher drive level. Unfortunately, marketing usually spec's the inductance at it's maximum level (at a high drive level.)
8H per kohm of reflected load is a rule I've heard more than once (or 45H for 5.6K.) 8H/kohm gives us a -3 dB point of 20 Hz. You can correct for low primary inductance a little by adding more bias current. Add enough current so that the load line biased up for a 2.6K load.
2.46K = 1/(1/Xl)^2 + (1/RL)^2 = 1/[1/(2*PI*20*23H)^2 + 1/5.6k^2]
Besides the primary inductance, keep an eye on the leakage inductance and total parasitic capacitance on the primary side. Usually, not always, the secondary side's capacitance is low enough not to be an issue.
If the primary inductance is inadequate, audio current from the output valve will be diverted from the load into that inductance. In other words, distortion will rise at low frequencies. As a rough rule, the reactance of the primary inductance ought to be equal to the reflected load resistance at the lowest frequency of interest. Thus, I would suggest that for 20Hz, you would like to see 45H. (Which is exactly what VoltSecond suggests.)
I got an answer from Lundahl today, they can make a pair of LL1688 gapped for 85mA without extra cost for the modification.
I asked for 80-85mA airgap but 80 was too close to 70 according to Mr Lundahl himself.
I guess the inductance will drop to maybe 50H from the original 70, but I think that will be good enough for practical purposes.
I asked for 80-85mA airgap but 80 was too close to 70 according to Mr Lundahl himself.
I guess the inductance will drop to maybe 50H from the original 70, but I think that will be good enough for practical purposes.
Try a Tribute transformers
I have ordred za pair of Tribute amorphous core tranformers for my GM70 project. You can specify any parameters you want. There is only a delivery time issue. Transformers sounds very good. Lundahl is good but a bit less civilized and dynamic (I have compared it to LL1623/160mA - not ideal for this aplication but you can taste it's tone)
www.tribute-audio.nl
regards
Maciej
I have ordred za pair of Tribute amorphous core tranformers for my GM70 project. You can specify any parameters you want. There is only a delivery time issue. Transformers sounds very good. Lundahl is good but a bit less civilized and dynamic (I have compared it to LL1623/160mA - not ideal for this aplication but you can taste it's tone)
www.tribute-audio.nl
regards
Maciej
How much does a pair of suitable Tribute transformer cost?
One major reason to buy from Lundahl is that they are made here in Sweden so payment and shipping will be a piece of cake.
One major reason to buy from Lundahl is that they are made here in Sweden so payment and shipping will be a piece of cake.
it was about 600Euro for a pair, 10kg each transformer, potted, about 50Euro for sending them to Poland
Maciek
Maciek
- Status
- Not open for further replies.