Hi David A couple more points 1) Add say a 33 ohm resistor in series with the op-amp output driver U7 . The transformer and associated load may be too much for a little opamp, so best to isolate it from a load that may cause it to become unstable- with the series resistor. It might not be a problem with a slow opamp like that, but if a faster opamp was substituted, there could be problems.
2) Yeah the 100uH inductor is handsome but there are side effects. If it is there to stop rubbish getting in and out, consider that a 100uH inductor might not look very inductive at high HF, low VHF. To be sure to to be sure, add a ferrite bead in series with it, or something that can be reasonably assured to still work as an inductor >10 MHz. say 1uH... 10uH + bead is probably a reasonable single inductor solution, also. The input will look like a short circuit at turn on, so the inductor charges up . The the capcs are all charged, the load goes away and of course the inductor wants the party to continue, hence the spike. A zener works OK depending on the energy in the inductor. The high voltage ring is usually quite high frequency and an electrolytic cap I find looks terrible and can do a pretty good ring suppression damping job. The zener is insurance. The HV ring will often go right through the switcher and to the CPU.... Yes increasing R21 will help but not as fast as a inductor will help. the series resistor R21, maybe a few ohms is usual to also stock HF ringing. 3) Ceramics. Alot of switchers now assume you are using ceramics. SO there is often that assumption in their control loops. Of course, as you know, the ceramics with their 'zero' ESR, get in on the act of ringing with the stray inductances around the place, and there are many of those. The electro in parallel would be hills those rings and HF components. Crappy electros see are good for something. The tantalum is good but the ESR is not as low , especially with the switchers heading up into the MHz. Tantalum is good for bulk capacitance for supplying the needs of the load, but not as good for dealing with the high ripple currents at high frequencies that will you likely get . And also, the ESR of a Tanalum varies ENORMOUSLY with quality and how much you pay. I have some 330uF 2V tantalums on my PCB for my SDR and I specify the mfrs part number to ensure I get precisely the part I need. regards On 29/09/2014 11:16 AM, David Rowe wrote: > Hi Glen, > > Thank you very much for your input. I will carefully consider each of > your suggested changes. > > At this stage I am feeling conservative - for the Betas I want to > minimise risk as any bugs we introduce from changes will results in 100 > repairs. > > So I'm inclined to separate mods to: > > (i) those that are low risk to add on the Betas > (ii) possible improvements we can try out in a later prototype stage. > > So the 100uH inductor kick is a good example of something we need to be > very careful with. Might be best to replace that with a bead, as > currently tested in the prototype. We could always increase R21 to form > a RC filter blocking switcher hash from re-radiating. > > Re (4) I could hear reduced switcher noise on the speaker output when I > tacked an electrolytic across C25, which was initially specified as a > ceramic (not sure what material however). Had assumed a tant would be > just as good? > > Cheers, > > David > > On 29/09/14 09:00, glen english wrote: >> I have some schematic comments >> >> Critical : >> >> 1) >> Try to avoid connecting input pins of the CPU to Vdd. This is bad ! >> >> IE PTT, SELECT, BACK (SW1,2,4) >> >> A latch up condition would mean destruction of the chip. >> >> If there needs to be input pins tied to Vdd, tie them through a low >> value resistor, like 470 ohms . NOT direct to supply. >> >> Improvements : >> >> 2) A stubber should be present on the switched mode supplu switching >> node. This will substantially reduce the ringing rubbish >> Jn C27/L2 >> >> 3) A 2 ohm resistor in series with C26 will also reduce radiated rubbish >> >> >> 4) A ceramic cap (X7R, X5R, X7S) would be a MUCH better choice for C25, >> C29, C28 in the power supply. >> Tantalums will under perform here . Eliminate C1. >> >> A bulk electrolytic may be a good choice on the input to reduce ringing >> and reduce Q of the equiv circuit. Put there where C1 was. >> >> >> 5) The 100uH inductor is a loaded gun....it may kick 2x or more voltage >> into the switched mode chip at power up- suggest the electrolytic be >> across C1, and/or a zener to clamp to spike >> >> >> comments >> >> 4) VREF+ may have too high a value resistor (R10) - successive ADC >> operations will deplete the reference stored in the C10 cap and generate >> errors. IE the first conversion is good, but subsequent conversions >> will show a higher conversion because The VREF+ will sag. >> This may not be a problem in this application- but it is if you are >> trying accurately to measure say a battery voltage. >> >> 5) Rather than the half rail supply splitting resistors everywhere for >> the op amp Vhalf-rail (VHR) , why not drop a opamp down and provide this >> voltage. this would eliminate 2 large capacitors and 2 resistors. >> >> >> >> On 29/09/2014 6:28 AM, David Rowe wrote: >>> Yes I linked to the prototype schematic, as that was what I was >>> debugging. The beta schematic is being worked on now. >>> >>> On 29/09/14 00:53, Steve wrote: >>>> The articles schematic link is old, I think this is the newest one: >>>> >>>> http://sourceforge.net/p/freetel/code/HEAD/tree/smartmic/SM1000-C/SCH-SM1000-C.pdf >>>> >>>> >>>> >>>> ------------------------------------------------------------------------------ >>>> Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer >>>> Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports >>>> Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper >>>> Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer >>>> http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk >>>> >>>> >>>> >>>> _______________________________________________ >>>> Freetel-codec2 mailing list >>>> [email protected] >>>> https://lists.sourceforge.net/lists/listinfo/freetel-codec2 >>>> >>> ------------------------------------------------------------------------------ >>> Meet PCI DSS 3.0 Compliance Requirements with EventLog Analyzer >>> Achieve PCI DSS 3.0 Compliant Status with Out-of-the-box PCI DSS Reports >>> Are you Audit-Ready for PCI DSS 3.0 Compliance? Download White paper >>> Comply to PCI DSS 3.0 Requirement 10 and 11.5 with EventLog Analyzer >>> http://pubads.g.doubleclick.net/gampad/clk?id=154622311&iu=/4140/ostg.clktrk >>> _______________________________________________ >>> Freetel-codec2 mailing list >>> [email protected] >>> https://lists.sourceforge.net/lists/listinfo/freetel-codec2 >>> > ------------------------------------------------------------------------------ > Slashdot TV. Videos for Nerds. Stuff that Matters. > http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk > _______________________________________________ > Freetel-codec2 mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/freetel-codec2 > -- - Glen English RF Communications and Electronics Engineer CORTEX RF & Pacific Media Technologies Pty Ltd ABN 40 075 532 008 PO Box 5231 Lyneham ACT 2602, Australia. au mobile : +61 (0)418 975077 ------------------------------------------------------------------------------ Slashdot TV. Videos for Nerds. Stuff that Matters. http://pubads.g.doubleclick.net/gampad/clk?id=160591471&iu=/4140/ostg.clktrk _______________________________________________ Freetel-codec2 mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freetel-codec2
