an experiment proposal

Discussions and experiences with user adjustable gaps.
Post Reply
PakProtector
Posts: 31
Joined: Wed May 04, 2005 2:36 am

an experiment proposal

Post by PakProtector »

Hey-Hey!!!,
I would like to try a PP output TX with an airgap. I'd like to use an existing coil geometry. In order to facilitate a fairly easy set of experiments, I would be comfortable trading power capacity in order to use higher perm/lower max flux density core material.

I should do some measurements on the primary L of the coil stacked with M6, in order to determine a target to set the initial gap for. There are still a few questions...since L is fairly coupled to signal magnitude, the bigger signals are going to make the L value increase. In trading a highly sloped and probably curved L v. signal plot I'd like to figure out where to have the two 'lines' cross. I'd predict the gapped core would have more L at low signal, and less at high signal than the ungapped core.

That the higher initial perm materials also seem to be lower HF loss will make seperating the two core performances.

Any suggestions?
cheers,
Douglas
Question Authority, disobey it only after understanding it, and when absolutely required...
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Re: an experiment proposal

Post by dave slagle »

PakProtector wrote:Hey-Hey!!!

I'd like to figure out where to have the two 'lines' cross.
i think i see where you are going, but have a hard time seeing the lines cross by simply adding an airgap. When you add air into the mix, perm goes down. It becomes more linear, but at a lower overall level.

dave
PakProtector
Posts: 31
Joined: Wed May 04, 2005 2:36 am

Post by PakProtector »

Hey-Hey!!!,
So...perm goes down, so I'd want to start with more, and gap it( eventually) to a value equalling 3.5%Si Iron ( as it would be replacing M6 in an existing design.

No sense trying to eval gap v. not gapped with two different coil geometry. Since the design in question is not running to the power limits of Si-Steel, I plan on using higher perm material( that has lower B_sat values ) and reduce power expectations.
cheers,
Douglas
Question Authority, disobey it only after understanding it, and when absolutely required...
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

So...perm goes down, so I'd want to start with more, and gap it( eventually) to a value equalling 3.5%Si Iron ( as it would be replacing M6 in an existing design.
Boy have I been beaten up in the past for that idea. I still think it is a very valid concept and the people who do not see that are too obsessed with a single detail that they miss the "big picture"

I'll still stand by my contention that the high initial perm of nickel is a non-linearity and in absolute terms minimally gapped nickel is LESS linear than other cores. If you fix the coil geometry and know your inductance needed then as you suggest you will need a larger gap in the nickel version. For lack of a better way to say it, you get the average between the air and the core. What people miss is the fact that air is linear and when you average it in. When the goal is linearity, nothing beats airgapped nickel.
No sense trying to eval gap v. not gapped with two different coil geometry. Since the design in question is not running to the power limits of Si-Steel, I plan on using higher perm material( that has lower B_sat values ) and reduce power expectations.
But don't you know that running M-6 at a fraction of it's Bsat increases linearity?

I can hook a scope up to my bridge and look at BH loops. If I could come up with a good way to save and present the data I'd do some family of curves. I'd guess a PC based scope is the answer, but I ain't got one of those :-(

I can test from 0-400Vac from 50-15Khz and the only thing stopping me from doing the whole family of curves is a sane methodology that gets me all of the data.

dave
PakProtector
Posts: 31
Joined: Wed May 04, 2005 2:36 am

Post by PakProtector »

One of my Tek 502A's has a camera mount...:)

So for an average of air and ____, what do you suggest? 49 Ni? I think I'm running higher wattage than 80 Ni would stand. I should measure power output in the amp's current state, I think it would be close.

How many watts per square inch can 80% Ni do at 20 cps?

clip:Boy have I been beaten up in the past for that idea.

yeah well...there's bone heads everywhere...some of them are teachable, and some seem to have a negative 'teachability'.
cheers,
Douglas
Question Authority, disobey it only after understanding it, and when absolutely required...
dave slagle
Posts: 2085
Joined: Sat Jan 22, 2005 3:54 am
Location: NYC
Contact:

Post by dave slagle »

PakProtector wrote: So for an average of air and ____, what do you suggest? 49 Ni? I think I'm running higher wattage than 80 Ni would stand. I should measure power output in the amp's current state, I think it would be close.
it may sound odd, but power is really not the issue here. sure knowing the power, you can work the numbers to figure everything else out, but to me that is just as convoluted as putting an impedance number on a trasformer.

simply put.... Power cannot saturate a core. If you take a typical power transformer and plug it into the wall with an open secondary and you will be just under saturation wiht zero power output. now short the secondary and you will be delivering tons of power yet the flux on the core will not change (first order). Yes the transformer will get hot and probably fail, but the failure will not be a result of saturation. The failure will be due to copper heating.
How many watts per square inch can 80% Ni do at 20 cps?
the answer to this question is a function of losses and other circuit parameters and has little to do with nickel. sure 80% nickel saturates at half the flux of M-19 so by extension you would need double the turns of half-sized wire to keep from saturating an equivalent sized core (read window area) but this just means the copper will limit the power. The core won't know the difference.

one could actually argue from the core POV, that in W per Sq Inch 80% nickel can handle more power since its losses are less :-)

The watts can easily be worked into volts knowing the other circuit parameters, then based on the number of turns the flux can be worked out.

dave
Post Reply