| HB9CVQ |
Rating:  |
2009-01-14 | |
| wow, but |
Time Owned: 3 to 6 months. |
The following operating- and small test-lab report is based on my Flex 5KC (Kontron 1.66 GHz DUO x86 Intel,F-S-B 533/667 MHz,Intel Graphic contr. GMA950, 1GB DDR2 RAM, 160GB,IEEE1394,USB 2.0,RS232C,10/100/1000Mbps Ethernet, HD mini ITX Computer with 2 DVI ports built in, XP pro), ATU, 2.RX, S/N: 1908-7018, Firmware 1.3.0.0, PowerSRD v1.16.1. , Driver 3.2.0 build 1556. My Sample rate is 96kHz, 2048 Buffer (one Flex recommendation). I always implement good external EMC measures to all my rigs, just in case.
Installation in my shack: Oct. 08. The service/support, both from USA and from EU HQ in DL is unique, outstanding and was needed in 3 failure cases since. Major Cons-points have already now been positively addressed by Flex and may be solved this year! Good cooperation.
First installation revealed no TX function (BNC ATU cable -external supplier to Flex- was incorrectly installed in the factory). After this was fixed in the shack the unit performed well under SDR 1.14.
In a snow storm, with static charges building up in my external 160 to 40m remote ATU- doublet 2x25m/up12m-, ESD killed one “electronic relay” in the RX path of the HRFIO board. The new board was swiftly shipped from Germany and before updated there to extra protection, which works fine ever since. I upgraded at this time to PowerSDR 1.16.1. This fixed, among others, several earlier (QRQ) CW shortcomings. The board calibration procedure is unique, all tests are basically internally software driven and well automated. Only the tolerances seem sometimes a little tight. You must run the test eventually several times to pass.
The 3. failure was caused by intermittent contact problems in the ATU, resulting in no TX and strongly attenuated RX. Bridging the ATU in the shack with a sort double shielded BNC cable solved all problems and improved shielding in the TX/RX path remarkably (status for all futher tests). This ATU degrades the Flex performance, because the in- and out-coaxes are introducing up to 5cm long pigtails in the shield to ground! Impedance and shielding issues do then exist inside the F5KC. Flex HQ EU offered a voucher for returning this ATU,nice.
I am running the F5KC with a Heil Proset plus -HC5-, ACOM 2000A @ max. 1kW out(HB9 limit) into a doublet or 3 el. Yagi. The field strength in the shack is typically < 7V/m. All cables to the antennas and to mains are fitted with RF-chokes/filters, there is 15MB/s WLAN. A large metallic ground plane system under each operating positions is used and the rigs are bonded to that plane with short studs.
My other rigs are Orion 2 TT and an old mod. IC 761, TL922 (over 10Y). The QSO focus is 160 to 10m, CW QRQ, contesting, rag chew CW, SSB and DXing.
Major test equipment used: Spectrum analyzer HP 8591E (TG), EMI current probes, EMI field sniffing probes, directional coupler (MC), signal generator HP 8647A, scope Tek 2465A, DL 1kW 50Ohm –60dB, power meter LP100A, Counter Racal DANA 1999 TXO. The purpose of this report is to add user info, which is not so easily found elsewhere. Trying also to be more quantitative, some test-lab data is added.
Pros:
· High quality shielded enclosure. Good EMC design
· Excellent documentation, operating + service manual / know-how base Internet supported
· Very good TRX performance under CW and SSB (Digital not checked yet) conditions; almost QSK (>50ms break-in). Even above 35 Wpm up to 60Wpm ok, if using only RX1. QSK works ok with F5KC solo, not with Acom 2ka, yet (see fix below).
Very smooth normal CW operation with CWX Memory. According to PARIS calibration, CWX DIT (1unit) to DAH(3units) ok, element pause (1unit) ok, letter pause (3 units) ok, word pause (7units) ok, but only if CWX Drop and Repeat Delay (ms) matches Speed WPM range chosen. Safe choice: 10Wpm (1000/1000ms), 20Wpm (600/600), 30Wpm (400/400), 40 Wpm (300/300ms) and up to 60Wpm (200/200ms).
The CW monitor signal (Audio) follows the TX RF output almost instantaniously in time domain. Even the set CW TX rise time can be well detected here. CW Paddle closing to TX RF output delay seems around 20 to 24 ms.
· Very clean TX CW signal (e.g. tr =2ms, 40Wpm, key mode A) with ACOM 2000A at 1kW out, via 70dB 50-Ohm power attenuator, on the scope or spectrum analyzer.
· Clean CW TX RF wave shape output (paddle-Schurr Profi 2), bouncing does exist on rear key (8) jack: DC (2,3V open/ 1.2mA short), new software handles debouncing better. For 1 kW (ACOM2kA) a drive level of about 25 to 50 W is sufficient. DIT to DAH ratio is 1:3 and ok.
· RX mostly superb, also in contesting with very strong signals near by; nice clear display of PAN Adapter (up to 80KHz BW viewing)
· Amazingly good copy of 40Wpm CW with BW 250/100Hz. Quartz filters start ringing here.
· After getting used to operating the unit just by screen (mouse/KB), TX/RX surface is ok.
· What a difference in the Band Scope to other rigs. You are no longer “blind”. You see what is happening,can record it, and the numerous RX BW filters really help your ears. In Panascope mode you can monitor CW signal shapes on RX1 in frequency and time domain. Scope Mode e.g. 300(us) however corresponds more to 30ms/div (one DIT at 40Wpm CW) for SR 96k/B 2048.
· Two identical receivers exist, phase locked, if needed! A very linear S-Meter calibration (typical accuracy within 1 to 2 dB in the HAM bands) in S-Units and also in dBm is implemented. Preamp in or out is automatically included in calibration/display. Small signal input VSWR (scalar quantity, vector analyzer was unavailable) of RX ports is typically better 20 dB RL (SWR 1.2). Worst case is 17m - without preamp on -12dB RL (SWR 1.7) and -with preamp on- 22 dB RL. This is lab instrument grade quality and makes the S-meter very precise.
· Ideally suited for monitoring bands and documenting/recording e.g. willful QRMing.
· Designed for SO2R, up to 3 Ant ports (SO239- delay set best to 0ms for Acom2KA internal delay around 10ms), RX1, RX2 directly accessible and as loop circuit e.g. for inserting a preselectors (RX1).
· Ant RX2 (-23dBm test injection) to Ant 1 coupling, Mode CW, BW 1kHz , AGC 100%, fast: 160m 72dB (S 5); 80m 84dB (S3) (96dB with preamp); 40/30 m with preamp 104dB (S0); 20 to 10m 97 to 86dB (S3) preamp on. Isolation between Port Ant.1-TX (33W) to RX2-in, 50Ohm terminated, gives 111dB / 160m, 117dB/80m down to 95dB/10m.
· RX Ant port (oscillator) leakage, conducted emissions, typically better –60dBm. Mostly best values with preamp on.
· Worst case desensitizing/ overloading/dynamic range (10m preamp always on in F5kC!): single, unmodulated QRM signal of 28010 MHz @ S9+60, und RX 10KHz away CW, BW 1KHz down to 50Hz, 100% AGC fast, results in only 72dB (S7) to 85 dB (S4) dynamic range. Considering however 15dB preamp defaul gain here, hints to a dynamic range elsewhere of 100dB (very good). S9+50 and S9+40 @BW 250Hz show S4 and S3)
· A 7.3 MHz S9+60/50/40dB signal causes desensitizing at 7.01 MHz of only S4 /S1/ S0, noise floor increase.
· DSP NR and NB are the best I ever had in a HAM TRX, very effective, wide-ranging and individually adjustable.
· Absolute frequency accuracy: 160, 20, 10m : +0.4, +3.6, +6.3 Hz (much better than in the Orion2 test 10m : –133Hz)
Cons:
· CPU Fan is acoustically noisy and disturbing
· 160m drive level control is not continuous around position 11
· Temperature increase in PA in contest @ only 35W out, already up to +60C , which is still green range under PowerSDR1.16.1, starting @ room temp. = 22 C. Why is the bias current still holding on for about 6s after key off?
· LDG Electronics ATU quality is poor (5cm coax shield wire, RF-part without matched strip lines-single layer. I will return it to Flex.
· CW (Iambic on)-almost QSK- is still problematic (but see first fixes below)
Testing for time sequencing even with a 4 trace scope is complex, because of unclear (jitter) trigger issues in the TX PTT output.
2. Rx muting causes unpleasant noise for fast break in (50ms). Very careful adjustment needed of AGC-T, preamp in /out and longer break in times (300 ms) may help.
Some in QSO QSK Hot Switching (ARC FAULT)happened in the Acom 2ka (300 to 1000W out)and could not be well reproduced in lab-testing. Jitter (up to > 7ms) is observed in the PTT (TX1,2,3) circuit.
------------------------Fixes-----------------
First EMI Field Probe sniffing investigations (Jan./12/09), internally, inside the F5KC, reveal uncontrolled RF TX-coupling from the HRFIO board via a flat 10 line ribbon cable to HTRX (RFIO SV2) board. Inserting a high permeable (4700) ferrite core (1.5 turns = 13uH- Jan./14/09) here helps now better keeping the RF in the upper chassis compartment. QSK operation is now feasable (60ms, 40Wpm) with the Acom 2000A for the first time. 80m, 1kW, just using one cable: Flex TX1 PTT out, no delay to Acom key-in. 160m is still worse than 80m . On 40 to 15m all is working fine too. Above, up to 10m, there is probably some capacitive crosstalk and more cores over some distance on the ribbon cable are needed.
The given ribbon cable, however, is too short for a higher number of turns around the core. CW keying has now rarely any hick-ups and the RX muting noise issues have also mostly improved. A longer ribbon cable and consequently more windings on the ferrite/or better more ferrite beads along the cable, is expected to trigger major problem fixes here. This might turn out as relatively easy and cost-effective field fix.
------------------------Fixes-----------------
There were initially irregular TX-RX transition QSK (after) pulses detectable in the RX audio output (noise) inbetween two Flex Tx and/or Power Amp active/on cycles. These pulses seemed RX2 on-related and somehow speed pronounced/dependent (e.g. @ 10Wpm). There also is some interaction between speed and break-in settings.
· Internal key (Key (8) input) issue(prior to fix): above 36Wpm, (break in 50 to 300 ms) CW in monitor does not sound perfect, (perfect with CWX -200ms, 200ms- or with key board active). These effects are independent of SR/Buffers setting. Even with external stn RX used as monitor, the internal key does not feel very good (CW op preferences). Disabling the Monitor only solves the noise RX 1 problem, not in RX2. I will also try other keying methods next.
· The internal monitor (-80dBc: 20dB+ 60dB?) causes somehow a bit strange looking signals in TX mode (stray coupling?), but much better now without ATU in place. If 60dB are realized in one step in a small print area, unshielded, this could lead to some cross talk and stray effects.
· RX does suffer from phantom signals (e.g. 17m) in EU, caused by strong 41, 49m BC Stations. S9+60 in BC bands causes a noise increase (of over 15dB) and spurs up to S7 from 40m to 10m. Flex HQ Germany started working on an internal retrofit solution for EU, coming out this year)
· 6.13395 MHz @ -23dBm causes 18.084 MHz to be S5. I think one solution could be adding suitable HP-Filters, or a preselector inserted in the RX loop path, may be switched in frequency, by Flexwire (retrofit).
· Panadapter Scope Time Mode incorrect calibration. Divided displayed time in (us) by 10 = ms (400->40ms/div) for SR=96/B=2048. Use CW-DIT as calibration marker (Wpm 60=20ms, 40=30ms, 20=60ms).
· Using F5KC in general RX mode suffers from 2. RX S-Meter off by up to 8dB. Individual image rejection retuning is needed. Spurious phantom signals, without preselector, do exist.
· Internal computer is not really (extremely) powerful -due to EMC considerations?-, but the price difference to F5KA is substantial. So why buying F5KC at all? Well, all is integrated in one box, therefore probably less (external)EMI trouble, no discussion about the suitability of the chosen computer.
· Running N1MM contest logger with Windows XP on that internal CPU, keying the rig, 2.RX on, fine resolution of band scope on, DX-cluster on, can bring internal CPU usage up to over 70%. Stuttering CW is a known phenomenon in XP, as a non real time system. Therefore I will try WinKey next.
As a summary I can state, keeping the Orion 2 is a nice backup for the new, beloved Flex 5000C. F5KC is now mostly the prime radio in my shack. I tested the rig in two small CW contests, adjusted to new op-skills with 2 large 22” DVI monitors, and conclude: Very high potential.
I understand SDR means introducing new technology. I do not mind to be experimenting sometimes and help pioneering in a friendly community. It is surely educational and fun to learn.
It might be interesting in the future also to check out competitive SDR systems with direct AD conversion right after the antenna band pass filtering,for comparisons.
Andy HB9CVQ
|
|
