• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Ceramic Filter Selection

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I've posted about FM tuner design considerations previously. Please refer to the FM MPX spectrum graphic. Notice that stuff above 53 KHz. is of no interest. In the past, what could be found there was not much of an issue. That is no longer the case. HD digital trash is in that area and needs to be kept out of our analog circuitry.

A ceramic filter between the IF strip and the discriminator/MPX circuitry could be used to block the digital garbage. I'd appreciate suggestions for selecting a suitable bandwidth part. Too narrow a bandwidth and analog audio info. will be blocked. :( Too wide a bandwidth and the digital junk will not be blocked.
 
Last edited:
Do not know too much about the spectrum of the HD radio signal but I think you are more concerned about filtering out the HD stuff after FM detection and not in the IF stage. I guess the assumption is that the HD info is screwing up the signal in the stereo decoder is a possibility. If you are using a stereo decoder, then you have to pass the stereo information, 19KHz pilot and the 38KHz sidebands, if you want that to work. So a multi-pole LPF of 38KHz+15KHz = 53Khz before the FM detection device, I think should do the trick.
 
Hi guys,
I am not sure of exactly what the OP is asking either, I assume HD interference, of some sorts.
I see in the Pioneer SX-1250/1050 designs, the use of birdie filters between the FM detector (F6,F7) and the stereo PLL chip (HA1196), what are their purpose and what are the filter characteristics? In the mean time , I'l do some investigation.
Rick
 
Ceramic filters are used at IF. Block LC filters (often known as birdie filters) can be used between discriminator and stereo decoder. Not sure which the OP is asking about.

IF filters need to be around 280kHz for good stereo, but 220-250kHz is OK if narrow bandwidth has to win over distortion. 180kHz for mono or very uncritical stereo.

The plan is to employ transformer coupling in the IF strip. A ceramic filter will be used only between the IF strip and the FM detection setup. HD garbage is known to cause grief and I want to "nail" it at an early point. As can be seen from this page, the available parts point towards a 280 KHz. bandwidth ceramic filter. Therefore, some additional filtering will be needed between the FM detector and the MPX decoding circuitry, to suppress HD trash.
 
Hi Eli,

I am up to learning as you go along, can you post some schematics of your design, are you modifying an existing design or doing up a new one?
I just last week acquired a Crown FM1, the GI controller is dead, none to be found, but I got it going enough (opened up the VCO control loop) to do an experiment comparing my Si4735 design to the FM1, guess who won the test? I see many ways to improve the FM1, other than re-design the PLL control but wonder if it will ever work as good as the SiLab chips. Must realize they designed their chips knowing the interference issues with HD carriers. Have to move with the times!!
As long as you realize that you are making a lot of work and may never achieve the performance of a cheap chip like Si4735 or better Si4770.

Cheers
Rick
 
Hi Eli,

I am up to learning as you go along, can you post some schematics of your design, are you modifying an existing design or doing up a new one?
I just last week acquired a Crown FM1, the GI controller is dead, none to be found, but I got it going enough (opened up the VCO control loop) to do an experiment comparing my Si4735 design to the FM1, guess who won the test? I see many ways to improve the FM1, other than re-design the PLL control but wonder if it will ever work as good as the SiLab chips. Must realize they designed their chips knowing the interference issues with HD carriers. Have to move with the times!!
As long as you realize that you are making a lot of work and may never achieve the performance of a cheap chip like Si4735 or better Si4770.

Cheers
Rick


Rick,

This thread and some other threads I've been in are about DIYing a tubed FM tuner, from scratch. Modern integrated circuits may be fine, but they don't fit here. Until I retire from gainful employment (hopefully, not too much longer) the project is, and will remain, in the talking stage. I have acquired the requisite 3 section air variable tuning capacitor. No other specific purchases have been made.

I'm leaning towards a gated beam combined discriminator/limiter, along with 4X 6AU6 IF stages. Quadrature demodulation of FM is available on "chips". I'm planning on doing it the original way. :D MPX decoding will be via a ring demodulator built around a 6JU8 quad diode.

It seems I have to come up with a "birdie" filter, as described by DF96.
 
Last edited:
Note that purchased birdie filters, if you can find any, are likely to require SS-friendly terminations around 4.7k. This may constrain the valve stage driving it.

As an alternative, you might consider using one or more notch filters at the HD carrier frequencies (I am showing my ignorance of HD here, as we don't have it in Europe - does it have carriers?). That might be easier to design than a high-performance low pass. You could add a more gradual low-pass too.
 
Hi guys,

Eli, thanks for the clarification of your tubed FM tuner project requirements.
Any particular reason why? Let me guess, you love the tube sound? or the challenge :)
Designing a hi-performance tube tuner is certainly a tough challenge at least it would be for me.
I am unfamiliar with gated beam combined discriminator/limiter, so I will look it up out of interest from your supplied link.
If you are interested in a var cap with 5 gangs, then look on ebay for this:
Pioneer SX 1250 AWB 023 FM Front END AND CAP Tuner | eBay
I was looking at the LA3450 stereo decoder spec, that describes the incorporation of a birdie filters to reject 115KHz.
If you need parts you can look at older SS tuners such as Pioneer SX-1050(4-gang var cap) which has some anti-birdie filters between the discriminator and the stereo decoder. I have one of these as as a spares unit.
 
Hi guys,

Eli, thanks for the clarification of your tubed FM tuner project requirements.
Any particular reason why? Let me guess, you love the tube sound? or the challenge :)
Designing a hi-performance tube tuner is certainly a tough challenge at least it would be for me.
I am unfamiliar with gated beam combined discriminator/limiter, so I will look it up out of interest from your supplied link.
If you are interested in a var cap with 5 gangs, then look on ebay for this:
Pioneer SX 1250 AWB 023 FM Front END AND CAP Tuner | eBay
I was looking at the LA3450 stereo decoder spec, that describes the incorporation of a birdie filters to reject 115KHz.
If you need parts you can look at older SS tuners such as Pioneer SX-1050(4-gang var cap) which has some anti-birdie filters between the discriminator and the stereo decoder. I have one of these as as a spares unit.

Rick, Dr. Adler's seminal paper about gated beam tubes is here. Maybe I should throw the towel in and use something like the LA3450 for MPX decoding. IIRC that 115 KHz. is associated with HD2.

In any event, my plan is for a 6922 in cascode as the TRF front end. A 6U8 or similar pentode/triode will take care of the mixer and local oscillator requirements. I'm thinking that a 6AF6 dual "magic eye" can take care of signal strength and tuning accuracy indications. If it proves necessary to add AGC, 1 or 2 of the 4X IF stages can use the 6BZ6, instead of the 6AU6.

FYI, I acquired the NOS air variable tuning cap. from Fair Radio, at a decent price. Look here.
 
And a varactor based tuner is not to be used, because? the nostalgia aspect? manual only.
Seems to me like the IF section makes or breaks a great FM tuner.
Also i am to think that traditional IF stages using transformer/tuned circuits, perform better than Ceramic filters do, maybe not SAW types, but they are not to be found.
AGC in the IF and none in the TRF stage? I guess tubes have higher dynamic range than say a jfet,MSOFET,bjt types?
I read http://frank.pocnet.net/sheets/093/6/6BN6.pdf,
seems to be a rather simple circuit. I wonder how well it performs (THD?). If I was smart enough you might even be able to simulate the design.
From what I have read, detected audio is the order of volts.
I wonder if you can use a double tuned circuit for the quad coil and it will improve THD just like it does for a IC version such as HA1137? I do not know enough about this stuff.
Interesting none the less...
 
Also i am to think that traditional IF stages using transformer/tuned circuits, perform better than Ceramic filters do, maybe not SAW types, but they are not to be found.
AGC in the IF and none in the TRF stage?

Rick,

Much of the superior sound tubed tuners exhibit can be traced to the transformer coupled IF strip. Unfortunately, FETs will oscillate in a transformer coupled IF strip. That leaves a designer stuck with 100% ceramic filters and their sonic liabilities, if FETs are to be present in an IF strip. :( IF transformers are available for BJTs and those can be adapted to tubed circuitry via parafeed topology. Mouser carries Xicon brand parts. Xicon part #42IF129-RC is what would be used. Also, a fellow in Germany manufactures IF transformers designed, from the ground up, to work with tubes.

In an FM only setup, clipping within the IF strip is desirable. It's called limiting. ;) I'll look at AGC, if and only if info. gets lost.
 
Eli,

I do not understand why a jfet would oscillate and not a bjt? in the transformer coupled IF strip. Must be a feedback mechanizm?
I do understand why you wan to limit the IF signal in a FM radio, AM rejection ratio?
I know you want the discrete frequency delays to match thus group delay? I guess this some sort of phase distortion.
And the reason why you do not want to use a tradition ratio detector?
Parafeed topology, now you got me? Looking that one up. Okay an RFC as the load(higher impedance, higher gain), cap coupled to the load being say another transformer.
I know when I test tuners, I have situations where there are two stronger adjacent carriers. The Si4735 deals with this issue better than a traditional design such as the Crown FM1. What would you do in your design to get the same performance as the Si4735? and deal with the strong adj carrier issue. I think it is the very sharp IF filters and maybe something to do with the detection method.
I read in Crown FM2 that they used a second IF freq and used a Hitachi pulse count detector IC, which know very little about,, no specs. But the spec differences between Crown FM1 and FM2 are about the same.
Hope you do not mind me asking lots of ?, not too many experienced RF designers out there. I know I am not one of them, still do not know anything about scatter parameters. I guess I should have asked the EE at Motorola Comm when I worked there years ago. I recall a few very smart EE's, still have those designs using J310's and other stuff like helical filters.
 
Rick,

While ratio detectors are self limiting and Foster/Seeley discriminators are not, the O/P signal level of a ratio detector is only 1/2 that of a F/S discriminator. Both of the dual diode setups require a costly and (perhaps) unobtainable transformer between the IF strip and the demodulator. With a gated beam setup, you just shove the I/P signal in. Look here.

An early tube oscillator setup was the tuned grid/tuned plate topology. It seems FETs are more prone than triodes towards oscillation, when tank circuits are present in both the I/P and O/P circuitry. I learned that tidbit towards the end of the 1970s from a book called FET Circuits.

The parafeed idea has even more gain than that provided by RFC anode loading. The tube's anode is cap. coupled to the untuned winding of the BJT IF trafo. The tuned winding is connected (end to end) to the grid circuitry of the next stage. Bada bing, bada bam, bada boom - instant step up trafo. :D The damned low I/P impedance of common emitter BJT setups forces the coupling trafo to be used in step down mode. :mad: BTW, notice that the coupling cap. isolates the BJT IF trafo from the B+ rail, which would "fry" it.
 
FET IFs were rare. Most use a cascade of BJT LTPs in a chip. This gives good limiting and a reasonable signal strength indicator. Ceramic filters can be better than an IFT in the passband, but have poor rejection of out of band signals so the best option is to use both: ceramics for setting the IF bandwidth and IFTs for rejecting other stuff. Multipole LC filters were best, but could be expensive and lossy.
 
Ceramic filters can be better than an IFT in the passband, but have poor rejection of out of band signals so the best option is to use both: ceramics for setting the IF bandwidth and IFTs for rejecting other stuff.

So, my idea of using a ceramic filter between the transformer coupled IF strip and a gated beam discriminator actually has some technical merit. :) Thank you!
 
Bear in mind that ceramic filters may become non-linear if too much signal is applied. I think I would use them earlier in the IF, and use an IFT at the output. Two ceramic filters should be enough to set the bandwidth. You could make one of them switchable (I used diodes) to vary the bandwidth: I had a 280kHz, and then in the next stage a switchable 280 or 180kHz (or was it 220kHz - it was a while ago!).
 
I thought "birdies" was a reference to adjacent channel interference.
Alternate channel interference is not usually a problem.

Filtering of the recovered signal leaves a choice of Mono, or Stereo, or Stereo plus data, depending on the passband of the filter.
Leaving data in the recovered signal does not usually result in interference, the existing filters to separate the three bands are usually good enough.

Any added filtering will need to be steep to keep the Stereo passband and reject much of the data band.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.