Forums

Ferrite transformer query.

Started by Clive Arthur November 6, 2017
I use RM8 ferrite cores as signal transformers in the range 10kHz to 
100kHz.  It's a half duplex system so is switching between transmit and 
receive.

There's quite a lot of variability in the audible noise from these, some 
you can barely hear, others there's a pronounced clicking as transmit 
starts.

These have been hand wound by me and others for prototypes.  Would a 
'proper' transformer manufacturer pot or encapsulate these by default?

The cores are ungapped and very sensitive to assembly - measure the 
inductance, take it apart and re-assemble and you get a different 
inductance.  Is there some recommended way of putting these together, 
for example cleaning the faces with something, or using some sort of glue?

I don't want a toroid, thanks.

Cheers
-- 
Clive
On a sunny day (Mon, 6 Nov 2017 12:30:56 +0000) it happened Clive Arthur
<cliveta@nowaytoday.co.uk> wrote in <otpkm2$udj$1@dont-email.me>:

>I use RM8 ferrite cores as signal transformers in the range 10kHz to >100kHz. It's a half duplex system so is switching between transmit and >receive. > >There's quite a lot of variability in the audible noise from these, some >you can barely hear, others there's a pronounced clicking as transmit >starts. > >These have been hand wound by me and others for prototypes. Would a >'proper' transformer manufacturer pot or encapsulate these by default? > >The cores are ungapped and very sensitive to assembly - measure the >inductance, take it apart and re-assemble and you get a different >inductance. Is there some recommended way of putting these together, >for example cleaning the faces with something, or using some sort of glue? > >I don't want a toroid, thanks.
Is there not some spring loading mechanism to really push the halves together?
On 06/11/2017 13:15, Jan Panteltje wrote:
> On a sunny day (Mon, 6 Nov 2017 12:30:56 +0000) it happened Clive Arthur > <cliveta@nowaytoday.co.uk> wrote in <otpkm2$udj$1@dont-email.me>: > >> I use RM8 ferrite cores as signal transformers in the range 10kHz to >> 100kHz. It's a half duplex system so is switching between transmit and >> receive. >> >> There's quite a lot of variability in the audible noise from these, some >> you can barely hear, others there's a pronounced clicking as transmit >> starts. >> >> These have been hand wound by me and others for prototypes. Would a >> 'proper' transformer manufacturer pot or encapsulate these by default? >> >> The cores are ungapped and very sensitive to assembly - measure the >> inductance, take it apart and re-assemble and you get a different >> inductance. Is there some recommended way of putting these together, >> for example cleaning the faces with something, or using some sort of glue? >> >> I don't want a toroid, thanks. > > Is there not some spring loading mechanism to really push the halves together? >
Yes, we are using the appropriate spring clips. Cheers -- Clive
On Monday, November 6, 2017 at 11:31:03 PM UTC+11, Clive Arthur wrote:
> I use RM8 ferrite cores as signal transformers in the range 10kHz to > 100kHz. It's a half duplex system so is switching between transmit and > receive. > > There's quite a lot of variability in the audible noise from these, some > you can barely hear, others there's a pronounced clicking as transmit > starts. > > These have been hand wound by me and others for prototypes. Would a > 'proper' transformer manufacturer pot or encapsulate these by default? > > The cores are ungapped and very sensitive to assembly - measure the > inductance, take it apart and re-assemble and you get a different > inductance. Is there some recommended way of putting these together, > for example cleaning the faces with something, or using some sort of glue? > > I don't want a toroid, thanks.
https://en.tdk.eu/inf/80/db/fer/rm_8.pdf Is the data sheet for the EPCOS RM8 cores and accessories. It includes a pair of spring clamps that do hold the core halves together quite tightly. It does make sense to wipe the mating faces clean with a low viscosity solvent. If you buy gapped cores, the gap is only about 0.2 mm which drops the inductance by about a factor of ten, and a ten fold smaller inadvertent gap would be enough to make a significant difference. -- Bill Sloman, Sydney
On Mon, 6 Nov 2017 12:30:56 +0000, Clive Arthur
<cliveta@nowaytoday.co.uk> wrote:

>I use RM8 ferrite cores as signal transformers in the range 10kHz to >100kHz. It's a half duplex system so is switching between transmit and >receive. > >There's quite a lot of variability in the audible noise from these, some >you can barely hear, others there's a pronounced clicking as transmit >starts. > >These have been hand wound by me and others for prototypes. Would a >'proper' transformer manufacturer pot or encapsulate these by default? > >The cores are ungapped and very sensitive to assembly - measure the >inductance, take it apart and re-assemble and you get a different >inductance. Is there some recommended way of putting these together, >for example cleaning the faces with something, or using some sort of glue? > >I don't want a toroid, thanks. > >Cheers
After cleaning, rub the core half mating surfaces together to ~polish off any hard or insoluble residue, before installing the (clean) bobbin. RL
On Mon, 6 Nov 2017 12:30:56 +0000, Clive Arthur
<cliveta@nowaytoday.co.uk> wrote:

>I use RM8 ferrite cores as signal transformers in the range 10kHz to >100kHz. It's a half duplex system so is switching between transmit and >receive. > >There's quite a lot of variability in the audible noise from these, some >you can barely hear, others there's a pronounced clicking as transmit >starts. > >These have been hand wound by me and others for prototypes. Would a >'proper' transformer manufacturer pot or encapsulate these by default? > >The cores are ungapped and very sensitive to assembly - measure the >inductance, take it apart and re-assemble and you get a different >inductance. Is there some recommended way of putting these together, >for example cleaning the faces with something, or using some sort of glue? > >I don't want a toroid, thanks. > >Cheers
What is the inductance? How much does it change? Does it matter? -- John Larkin Highland Technology, Inc lunatic fringe electronics
Are you using DC bias?  Is the air gap consistent?

You can try to arrange the circuit to prevent DC bias or startup transients. 
Using a push-pull rather than single-ended circuit, for instance.

Ferrite has nonzero magnetostriction, so it will necessarily vibrate a 
little, at whatever frequency you're magnetizing it at, including startup 
transients or DC bias (or twice the frequency, if unbiased).  This vibration 
can only be dampened with potting and [acoustic] shielding.

A good example is the old 15.7kHz CRT flyback, which sometimes sang terribly 
loud because of poor manufacture, but was always audible thanks to 
magnetostriction.  Well, maybe not audible to old people, but us 
not-so-young-anymore-'ins know this well.

Tim

-- 
Seven Transistor Labs, LLC
Electrical Engineering Consultation and Contract Design
Website: https://www.seventransistorlabs.com/

"Clive Arthur" <cliveta@nowaytoday.co.uk> wrote in message 
news:otpkm2$udj$1@dont-email.me...
>I use RM8 ferrite cores as signal transformers in the range 10kHz to >100kHz. It's a half duplex system so is switching between transmit and >receive. > > There's quite a lot of variability in the audible noise from these, some > you can barely hear, others there's a pronounced clicking as transmit > starts. > > These have been hand wound by me and others for prototypes. Would a > 'proper' transformer manufacturer pot or encapsulate these by default? > > The cores are ungapped and very sensitive to assembly - measure the > inductance, take it apart and re-assemble and you get a different > inductance. Is there some recommended way of putting these together, for > example cleaning the faces with something, or using some sort of glue? > > I don't want a toroid, thanks. > > Cheers > -- > Clive
On 11/6/2017 6:30 AM, Clive Arthur wrote:
> I use RM8 ferrite cores as signal transformers in the range 10kHz to > 100kHz.&nbsp; It's a half duplex system so is switching between transmit and > receive. > > There's quite a lot of variability in the audible noise from these, some > you can barely hear, others there's a pronounced clicking as transmit > starts. > > These have been hand wound by me and others for prototypes.&nbsp; Would a > 'proper' transformer manufacturer pot or encapsulate these by default? > > The cores are ungapped and very sensitive to assembly - measure the > inductance, take it apart and re-assemble and you get a different > inductance.&nbsp; Is there some recommended way of putting these together, > for example cleaning the faces with something, or using some sort of glue? > > I don't want a toroid, thanks. > > Cheers
Can you use a gapped core with more turns? I think you get more consistency with gapped cores. Mikek
Am 06.11.2017 um 20:31 schrieb amdx:

> &nbsp;Can you use a gapped core with more turns? > &nbsp;I think you get more consistency with gapped cores.
Yes, just a little bit of Kapton tape to create a small, repeatable air gap.
On 06/11/2017 19:31, amdx wrote:
> On 11/6/2017 6:30 AM, Clive Arthur wrote: >> I use RM8 ferrite cores as signal transformers in the range 10kHz to >> 100kHz.&nbsp; It's a half duplex system so is switching between transmit >> and receive. >> >> There's quite a lot of variability in the audible noise from these, >> some you can barely hear, others there's a pronounced clicking as >> transmit starts. >> >> These have been hand wound by me and others for prototypes.&nbsp; Would a >> 'proper' transformer manufacturer pot or encapsulate these by default? >> >> The cores are ungapped and very sensitive to assembly - measure the >> inductance, take it apart and re-assemble and you get a different >> inductance.&nbsp; Is there some recommended way of putting these together, >> for example cleaning the faces with something, or using some sort of >> glue? >> >> I don't want a toroid, thanks. >> >> Cheers > &nbsp;Can you use a gapped core with more turns? > &nbsp;I think you get more consistency with gapped cores. > > &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; Mikek
No, the bobbin is pretty full, so no more turns. Cheers -- Clive