38.9 V 39.5 Mhz TV IF.

[colly0410]

Why did the UK (& presumably Ireland) use 39.5 Mhz for their IF for 625 lines TV when most other countries used 38.9 Mhz? I understand South Africa uses 38.9 Mhz on system I..

[Terry]

38.9MHz was chosen by the CCIR long before the UK considered using the 625-line system but this created other problems due to the need for early 625-line sets to cater for both 625 and 405 line signals.

One suggestion is that the 39.5MHz figure came from Mullard:

Mullard at least on occasion advocated dual-standard IF strips whose basic bandpass characteristic was 6 dB down at 34.65 and 39.5 MHz respectively, and so which included the Nyquist slopes for both the 405- and 625-line cases ...

This is taken from Synchrodyne's post on the subject on the old VRAT forum. The original thread can be seen here.

[Synchrodyne]

I haven’t yet found the definitive reason for the choice of the 39.5 MHz.  As far as I can find, the reasoning behind BREMA’s decision were not recorded in the pages of either Wireless World or Practical Television.  I can’t say that I’m actively looking these days; rather it is a sleeper that I might just happen across when looking for something else.

My best guess remains that it was because for low cost dual-standard receivers, 39.5 MHz allowed a relatively simple double-Nyquist IF bandpass that did not ridiculously restrict the 625-line video bandwidth.  The 4.85 MHz that it allowed was hardly stellar, and perhaps it could be corrected to a bit over 5 MHz by a post-demodulator tilt circuit, although that was hardly likely on a low-cost receiver.  But it was probably visibly better than the paltry 4.25 MHz that would be obtained at 38.9 MHz with a double-Nyquist curve.  Receiver makers could, of course, for their less cost-constrained models, use largely separate IF-strip input bandpass filters and traps for each standard, in which case it wouldn’t matter much whether the 625 IF was 38.9 or 39.5 MHz.

The price to be paid was a steep increase in the image rejection requirement in order to accommodate the co-siting of transmitters at channels n and (n+10), which was part of the matrix agreed at the Stockholm 1961 ITU meeting, before the 39.5 MHz number was in play.  This required the addition of an input tuned circuit to the German-style of valved UHF tuner, which normally had an aperiodic input with a bandpass interstage.  That probably cost a dB or three in terms of noise factor, which had little reserve to begin with.  (As it happens there is a concurrent discussion about valved UHF tuners in this thread over at UKVRR:  http://www.vintage-radio.net/forum/showt...p?t=140359.)

39.5 MHz was probably about as high as the IF could be pushed without getting into impossible image rejection requirements.  Evidently it checked the boxes in the self-interference avoidance category.  And it placed the 625-line lower adjacent-channel sound at 41.5 MHz, which was the same as the channel B1 sound-trap, also a desirable feature at the IF strip entry given the high channel B1 field strengths found in some parts of the UK.  So that was a small economy, and one that persisted into the single-standard receiver era for as long as the 405-networ remained live.


Cheers,

Steve

[colly0410]

I was thinking something to do with UK being a UHF only country, then I remembered Ireland used system I at VHF. As Synchrodyne says don't suppose it matters in single standard sets what IF is used, but did UK set makers carry on using 39.5 in SS days? I have an NEI TV from 1987 that still works OK, wondering what IF that uses? Just noticed the valve UHF tuner thread, will have a read of it later, thanks for info...

[Synchrodyne]

As far as I know, the UK setmakers continued to use the 39.5 MHz in the single-standard era.  For as long as the legacy 405-line transmissions continued, there was an advantage in that the 41.5 MHz trap did double duty as the lower adjacent channel sound trap and the channel B1 sound trap.  A full range of SAWFs was developed around the 39.5 MHz IF in the late 1970s, with quasi-split sound variants added a bit later, and the later still, SAWFs that accommodated NICAM sound.  So, there was no compelling reason to change away from 39.5 MHz.

I have a vague notion that one or two of the European setmakers may have used the 38.9 MHz IF in the 1980s or so, but right now I cannot find a reference.

Returning to the double-Nyquist notion, I first thought of that when reading this Wireless World (WW) 1961 October item about the Pye prototype dual-standard receiver:

   

Clearly, the basic IF bandpass curve was of the double-Nyquist type, with -6 dB points at 34.65 and 38.9 MHz.  The narrower bandpass required for 405 was obtained simply by means of the 38.15 MHz sound trap.  As shown, this gave a 625 video response of 4.25 MHz at the -6 dB point.  That would have been poor even for System B, where one would hope for at least 4.5 MHz and a bit beyond.  The 405 video bandwidth I’d estimate as having been around 2.75 MHz.  It’s hard to tell from the graph, but the lower-side Nyquist slope should have been a bit steeper than the upper-side slope.  Given that the lower-side trap was 1.75 MHz was below the lower -6 dB point, and the upper-side trap 2.0 MHz above the upper -6 dB point, it seems likely that the lower-side slope would have been a bit steeper.  

Anyway, given that the 405 IF was more-or-less cast in stone, whereas that for 625 was not, a way to improve the 625 video response was by moving the IF above 38.9 MHz.  One may imagine that some of the setmakers would have lobbied BREMA to do this, so that at least they had the option of using the double-Nyquist IF curve for their cheaper receivers.  And as said previously, 39.5 MHz was probably about as high the IF could be set without running into impossible image rejection requirements and without overlapping the lower edge of Band I.  It helped a little but not that much, and I imagine that many dual-standard receivers did not use the dual-Nyquist approach in the interests of obtaining a video bandwidth that was towards 5.5, and at least say above 5 MHz.

Here is the commentary on the image rejection requirements from WW 1962 October:

   

From the opening statement of that item, I should expect that the BREMA decision in favour of 39.5 MHz to have been announced in a preceding WW issue, perhaps with reasons, but if so, I cannot find it.


Cheers,

Steve

[colly0410]

Was lucky that there were no broadcast stations in the 30 to 40 Mhz spectrum, that would have messed things up big time. When I was a squadie we'd use 36 to 38 Mhz for artillery - inf/armour/cav coms, what effect that had on local TV's I never knew. I got into trouble for tuning a radio to 41.5 Mhz to listen to TOTP's in Aldershot area, was a bit distorted as our sets were FM & TV sound was AM..

[Terry]

I don't think anything used anything over 30MHz when TV was being developed in the 30's but, on the other hand, the early sets were mainly TRFs, anyway.

But the earliest superhets that I encountered had IFs of 16 & 19.5MHz. Never knew of any examples of SW breakthrough, though ...

[colly0410]

I suppose the none UK set makers made the same TV's for sale in the UK & South Africa & just left out the VHF tuner for UK bound sets. I presume they'd use 38.9 Mhz IF so they could use the same production lines as system B/G sets & just up the sound to 6 Mhz for system I sets..

[Synchrodyne]

Was lucky that there were no broadcast stations in the 30 to 40 Mhz  spectrum, that would have messed things up big time. When I was a squadie we'd use 36 to 38 Mhz for artillery - inf/armour/cav coms, what effect that had on local TV's I never knew. I got into trouble for tuning a radio to 41.5 Mhz to listen to TOTP's in Aldershot area,  was a bit distorted as our sets were FM & TV sound was AM..

In most cases TV IF channels were chosen to be in a relatively quiet part of the spectrum, and once assigned as a national standard, whether de jure or de facto, were usually given some measure of protection against future deleterious occupation.

In the US case, the 40 MHz “high” IF channel was developed by a subcommittee of an industry body, namely the RMA, in 1948, but no doubt watched by the FCC.  The FCC then used it as the basis for its UHF channel allocation plan, so had a vested interest in keeping it clear of interference.

In the UK, the standard TV IFs were also developed by an industry body, namely BREMA.  Possibly the GPO was represented on the pertinent subcommittee, but would at least have had a watching brief.

Returning to the South African choice of 38.9 MHz for System I, as well as easing the UHF image rejection requirement, it may also have been seen as having a slight advantage in respect of local oscillator placement in Band III.

For example, the local oscillator frequencies for channel ZA4 were 214.15 MHz for the 38.9 MHz case, and 214.75 MHz for the 39.5 MHz case.    Channel ZA9 vision was at 215.25 MHz, with a lower vestigial sideband extending down to 214.0 MHz.  The 214.75 MHz local oscillator frequency (for 39.5 MHz IF) would be around 8 dB down on the receiver Nyquist slope, the 214.15 MHz local oscillator (for 38.9 MHz IF) around 24 dB down, and so less likely to cause any interference problems.  So, absent any compelling reason to choose 39.5 MHz, 38.9 MHz looked to be the logical choice.


Cheers,

Steve