(30-01-2016, 11:09 AM)ppppenguin Wrote: [ -> ]The output is harder. Again it's TF hybrid with a NE5532. I don't think you can do a balanced floating output without more components. Electronic balanced floating outputs are tricky things, with positive feedback to ensure that when you ground one leg the other leg doubles in amplitude. More likely it's balanced around ground but not floating. So you can't really ground one leg of the output. Easy enough to check this with a signal passing through. Just ground one leg and see what happens. Hence I'll be taking an unbalanced output between one leg and ground. This also halves the amplitude of the output so hope the gains presets have enough range to compensate.
Yes, most are non-floating types in my experience. Just an inverter taking a copy of the "hot" signal to produce the "cold". As you say, ignore the cold and take the hot, albeit at -6dB. Saves design effort - saves making sure there are no RFI injection issues.
You're right that grounding the cold or hot of a quasi-floating output will give the correct amplitude, albeit with a 6dB loss of headroom, which is totally academic here because the single-ended inputs have already lost 6dB headroom unless they are using 30-0-30 for their op-amps
. If you ground the cold at the remote end rather than at the router output end, you'll get the advantages of ground noise cancellation - which probably isn't an issue here as I think(?) the destinations are adjacent to the router, but is worth mentioning for long cable runs...
I've a nasty feeling that any gain tweak presets in the router might only be fine-tuning types - ±1dB or thereabouts. Presumably the destinations have input level tweaks? Not sure what you're using, but I'd expect a non-pro modulator to have an adjustment and some means of indicating the level - especially when it's expecting to be used with a wide range of sources?
Anyway, you're definitely getting there
Yes, the Auroras and UHF modulator have input level pots. Level indicator? Forget it. Wind it up until it distorts, then back off a bit. Fortunately the DVD player and freeview boxes have about the same audio output levels so no need to balance them up. The rails are +/-12V so not really full pro capability and headroom.
Pending checking out the audio router I've decided to go with the Invotron routers. I've designed a simple interface to interlocked pushbuttons. Using bits from the junkbox it has
4x 74LS30 8 input NAND gates (would have used HC30 or HCT30 if I'd had them) as a 16 to 4 bit encoder
3x 4051 8:1 analogue mux
1x 74HCT138 3:8 decoder
1x 4060 divider with clock osc
Plus a couple of resistor packs as pullups, some decoupler caps and not a lot else.
Drawing up the schematic properly forces me to think through the loose ends of the design as well as making it easier to assemble and to maintain afterwards. I wheeled out old fashioned Orcad for DOS and the keystrokes are slowly coming back to me. I can work faster in this than the newer Orcad for windows. When I've finished I'll convert the schematic to Orcad for Windows for ease of maintenance and printing.
First pass of controller schematic.
U1 (4060) is an oscillator and divider. Q2 and Q3 sequence round the 4 switch banks via U2 (4051) which applies 0V to the common point of the relevant bank. U6-U9 (74LS30) encode up to 16 switches to 4 bit binary as needed by the router.
When the system is in standby mode U2 is disabled and either U3 (for off air) or U4 (for testcard) is enabled. These force the selection on each bank. No point in forcing the preview bank as the monitoring system is switched off in standby.
U5 (74HCT138) also decodes Q2 and Q3 to pulse the latch enables in the router. Using Q0 and Q1 to enable the decoder ensure that the latch enables are only pulsed when the data is stable.
I think I have chosen the arrangment of Q0 and Q1 so that no glitches occur on the '138 outputs, even though the counter outputs are not synchronous. Actually even if U1 was a synchronous counter you should never rely on the outputs being exactly cotimed. Get the design wrong and you can still get glitches, nasty very short ones that give horrible intermittent problems. This is not the time or place for a tutorial on synchronous logic design but I'm sure there's plenty of guidance online.
If anyone has done a detailed analysis of that schematic they will find that I've actually got the Q1 and Q0 inputs to the '138 the wrong way round to avoid glitches. I re-did the timing diagrams and worked it out again from scratch. Because Q1 always changes later than Q0 (it's a ripple counter) then the safe epoch is when Q1=0 and Q0=1. Q0=Q1=1 is also safe but that doesn't give much timing margin from when Q2 and Q3 can change.
Got the audio switcher on the bench. Establshed how the ins and outs are wired u which wasn't immediately obvious. Found one of the 4042 latches merrily singing away for no good reason. Swapped it with another to prove it was faulty. I have some spares in my junk box.
The designers have been rather naughty. The 4051 analogue switches are being run on +/-12V, that's 24V between VDD and VEE. Data sheet says abs max is 18V. Tut tut.
Here's the audio router being tested. Controls are a hex switch for data and croc clip lead for latch enable.
I only tested inputs 1-10 as that's all we'll be using and all I've fitted phonos for. You can see the phonos nicely bodged on to the interface PCBs. The top row of input phonos sticks up slightly above the 1U rack so will need a blank panel above it. Not a problem.
On the main router PCB you can see:
2 groups of 8 input buffers (left and right), nearest the edge connectors. We're only using the left channels. Good old mono.
2 sets of 8x 4051 8:1 mux to give 16x4 routers. I nicked the far right one to replace a duff one on the left channel router.
Beyond the 4051s are 2 sets of 4x 4042 quad latches. I nicked the far right one to replace a duff one.
Furthest away are the 2 sets of 4 output buffers.
The pairs of chips in the near corners are level shifters. They take 5V controls and translate them to the 12V needed for the 4042 and 4051. As I said earlier, the designers were very naughty using the 4051s on +/-12V when the abs max rating is 18V. I wonder how they justified that to themselves. You really need +/-12V for professional audio. Actually you really want more than that, +/-15V or even 18V if you're serious about dynamic range.
Should now go back to the video router and a full check on all crosspoints. Then I can get on with building the interface and button panel.
Run through the video router and it's fine. Some variablility in the black level between inputs. Won't matter as within reason the absolute DC level of a video signal is ignored by all video inputs. There's some kind of clamp or DC restorer on each output that may be getting things wrong. Not sure it's worth tracing.
More worrying are a number of bead tantalums. I'm tempted to just replace the lot as an insurance. 2x 22uF per output channel and 2x 4.7uF per regulator. The audio router also has tants for its regulators. For the sake of 10x 4.7u (I'd use 10u as I've loads of them) and 8x 22u it's probably worth the effort.
Decided to change direction again.
I've ordered a couple of these modules:
http://pldkit.com/xilinx/xm2c2 Just under £30 the pair including postage from Estonia.
I have some panels with sets of "Schadow" push switches which are just right for the job. The top row of 10 switches is for preview, the 3 sets of 6 are for BBC1, ITV, BBC2. i have an assortment of orange and black presses which are easily interchanged. Some have a slot for a label, some don't. I'll find a sensible looking set of presses.
I think that the CPLD module will happily do a multiplexed LED drive without any external buffers. So I'll fit a ribbon cable from one side to drive LEDs and buttons. Plus a ribbon from the other side to drive the router. Power at 5V will come from the router.
One slight unknown is whether the 3.3V outputs from the CPLD will drive the MC14504 level shifters on the routers. These need 5V CMOS levels which are only just about reached by 3.3V signals. The MC14504s can be programmed to work with TTL levels, simply by changing the voltage on pin 13. 3.3V CMOS will easily drive 5V TTL. This is likely the simplest solution. Alternatively I could lower the Vcc on the MC14504s to 3.3V. A bit more complex and not covered in the data sheet though it would surely work.
Another minor unknown is what sort of series resistors I will need to set the LED current. I'll be multiplexing 4 LEDs so need 4x normal LED currrent. The LEDs would typically be run at about 5mA so 20mA drive needed in a x4 mux. The CPLD outputs will deliver this easily, the LED voltage drop is likely to approach 2V at this sort of current so (3.3-2)/0.02 = 60R. So I'll start with 68R or 56R and see what happens.
The logic design will be some pretty simple VHDL. I have the Xilinx tools and have designed for this class of device before so I don't anticipate any problems. Correcting any design errors is a lot easier in VHDL than hardwired CMOS/TTL.
You can see mods I've made to the top switch panel. The pairs of BNCs and phonos are for 2 aux AV inputs. The toggle switch is to select off-air or testcards in standby mode. There will be a coaxial low voltage power socket on the back to take power from a wallwart. This will signal whether the system is running or in standby.
Today I wired up the control panel for the router. Found a decent set of key caps and repalced one faulty switch.
Then I worked out arrangments for the ribbon cables that connect the PLD module to the panel and the router. With luck the PLD module will arrive this week. I've started on the VHDL for it. Shouldn't be too complicated.
Need to make up the adaptor ribbon cable that goes from the router to a 20 pin ribbon cable plug. Also a short D15 to D15 cable to link the audio and video routers. That should complete the construction phase. Then it's debugging the controller - editing VHDL is easier than rewiring chips - and finding a suitable date to install it at the Musuem.